Sports Medicine 14 (5): 320-335. 1992 0112-1642/92/00 11-0320/$08.00/0 © Adis International Limited. All rights reserved. SPOll85

Running Injuries

A Review of the Epidemiological Literature

Willem van Mechelen

Department of Health Science, Faculty of Human Movement Sciences, Vrije Universteit en University of Amsterdam, Amsterdam, The Netherlands

Contents 32/ 322 324 324 324 325 325 326 326 327 327 327 328 328 328 329 329 329 329

330 330

33/ 33/ 33/ 332 332 333 333 334

Summary 1. Incidence of Running Injuries 2. Severity of Running Injuries 2.1 Medical Diagnosis 2.2 Site of Injury 2.3 Consequences of Running Injuries 3. Causes of Musculoskeletal Running Injuries 3.1 Characteristics of Runners 3.1.1 Age 3.1.2 Gender 3.1.3 Anthropometries 3.1.4 Malalignment 3.1.5 Muscular Imbalance, Restricted Range of Motion 3.1.6 Running Experience 3.1.7 Previous Injury 3.1.8 Participation in Other Sports 3.1. 9 Psychological and Behavioural Factors 3.2 Characteristics of Running Practice 3.2.1 Weekly Running Distance, Running Frequency and Weekly Running Time 3.2.2 Training Speed and Other Indicators of Intensity of Running Performance 3.2.3 Stability of Running Habits 3.2.4 Warm-Up, Stretching Exercises and Cooling Down 3.3 Characteristics of the Running Environment 3.3.1 Shoes 3.3.2 Running Surfaces and Hill Running 3.3.3 Time of Day and Season 4. Prevention Through Health Education 4.1 Complete Rehabilitation and Early Recognition of Symptoms of Overuse 4.2 Training Guidelines

Epidemiology of Running Injuries

Summary

321

Running is one of the most popular leisure sports activities. Next to its beneficial health effects, negative side effects in terms of sports injuries should also be recognised. Given the limitations of the studies it appears that for the average recreational runner, who is steadily training and who participates in a long distance run every now and then, the overall yearly incidence rate for running injuries vari~s between 37 and 56%. Depending on the specificity of the group of runners concerned (competitive athletes; average recreational joggers; boys and girls) and on different circumstances these rates vary. If incidence is calculated according to exposure of running time the incidence reported in the literature varies from 2.5 to 12.1 injuries per 1000 hours of running. Most running injuries are lower extremity injuries, with a predominance for the knee. About 50 to 75% of all running injuries appear to be overuse injuries due to the constant repetition of the same movement. Recurrence of running injuries is reported in 20 to 70% of the cases. From the epidemiological studies it can be concluded that running injuries lead to a reduction of training or training cessation in about 30 to 90% of all injuries, about 20 to 70% of all injuries lead to medical consultation or medical treatment and 0 to 5% result in absence from work. Aetiological factors associated with running injuries include previous injury, lack of running experience, running to compete and excessive weekly running distance. The association between running injuries and factors such as warm-up and stretching exercises, body height, malalignment, muscular imbalance, restricted range of motion, running frequency, level of performance, stability of running pattern, shoes and inshoe orthoses and running on I side of the road remains unclear or is backed by contradicting or scarce research findings. Significantly not associated with running injuries seem age, gender, body mass index, running hills, running on hard surfaces, participation in other sports, time of the year and time of the day. The prevention of sports injuries should focus on changes of behaviour by health education. Health education on running injuries should primarily focus on the importance of complete rehabilitation and the early recognition of symptoms of overuse, and on the provision of training guidelines.

The popularity of running as a form of exercise and recreation has grown rapidly since the I 970s, first in North America and later, from the beginning of the 1980s, in Europe. One of the reasons for this worldwide trend is the low cost involved: all that is needed is a tracksuit and a pair of running shoes. Running can be performed at any time and anywhere without constraints of a schedule. Reasons for jogging include health, fitness, pleasure, relaxation, competition and personal performance (Clough et al. 1989; Ooijendijk & Van Agt 1990). More and more people are taking part in major endurance events such as the New York, Los Angeles, Amsterdam or London marathons. In Switzerland the number of regular joggers has increased by 100% over the period between 1978 and 1984, being 8% of the total population (Marti et al. 1988). According to Walter et al. (1989) the number of Canadians jogging or running doubled from 15% in 1976 to 31 % in 1983. Jacobs and Berson

(1986) state that in the US there are 30 million runners of all levels, of whom 10 million run regularly and about 0.8 to 1 million enter races. In the Netherlands in 1978, 8% of the Dutch adult population was engaged in running, while in 1984 this number had increased to 17% (Manders & Kropman 1982, 1987). This 1984 figure meets the results of a telephone survey covering 1000 respondents which was carried out between mid-July and mid-August 1985 by the Amsterdam Market Research Institute Inter/View and which indicated that there are around 2 450 000 joggers in the Netherlands (in a total population of around 14 million). Ofthis total around 2 million run at least once a week. Knowing that, according to the 1990 statistics of the Dutch Sports Federation, soccer with about I million participants is the most popular organised sport it would thus appear that running is the major national sport in the Netherlands. Running is widely considered to have a preventive

322

Sports Medicine J4 (5) 1992

effect in respect of cardiovascular problems (Eichner 1983; Hartung 1980; Powell & Paffenbarger 1985). There are benefits associated with the 3 main risk factors in cardiovascular disease, namely obesity, hypertension and smoking (Eichner 1983; Koplan 1982). As a group runners smoke less, have a lower percentage of body fat and are less likely to have high blood pressure. Most runners and joggers are not members of any sports association, thus, do not appear in any membership records. Despite the lack of an organised framework many runners practise their sport intensively. Distances of over 90km each week are far from exceptional (H"Jlmich et al. 1989). The purpose of this article is to describe the incidence, severity, aetiology and preventive measures of running injuries. Prescribing prevention can only be done on the basis of knowledge of the aetiology of running injuries. Much of the existing information is presented as a result of the analysis of case series or

common sense. As explained by Walter et al. (1985) the methodology of case series however does not allow any conclusions on the causes of running injuries. However, epidemiological studies may allow these kind of conclusions. Therefore, this review of the literature is predominantly based on the outcome of epidemiological studies.

1. Incidence of Running Injuries There is no agreement in the literature regarding the incidence of running injuries. Different incidence figures are presented: person-incidence rates (number of injured runners per 100 runners), injury-incidence rates (number of injuries per 100 runners) and injury-incidence per exposure (number of injuries per 1000 hours of running). Differences in definitions of sports injury and sport participation, runner characteristics, research design and length of study may influence findings on injury incidence as described by Van Mechelen et al. (1990).

Table I. Incidence rates (%) of running injuries based on studies with an incidence time period of 1 year Reference

Koplan et al. (1982)

Blair et al. (1987) Lysholm & Wiklander (1987)

Effect on training and condition

No. of subjects, sex and average age (years)

Training load (km/week)

~ training,

693M,33 730F, 29

>10

37 38

40

24 68

medical treatment, medication ~ training > 7 days ~ training > 7 days

438M & F, 44 19 sprinters, 21 13 middle dist. runners, 19

Yzerman & Van Galen

~ training,

28 long dist. runners, 35 757M, 50W, 15-70

(1987)

medical treatment

15-70

Marti et al. (1988)

+training

4335M, 17-64

?

Incidence rate (%)

77 57 70

56 38

24

45.8

17-64 H01mich et al. (1989)

Stopped training

Clough et al. (1989)

Macera et al. (1989)

1310M,34

>30

31

489M marathoners, 33

1959 kmty

41

440 marathon drop-outs,

1212 km/y

49

39 37

52 49

~ training,

31 485M,42

medical treatment,

98W, 36

medication Walter et al. (1989)

~ training,

985M, 14-50+

49

49

303W, 14-50+

35

46

Ooijendijk & Van Agt (1990)

medical treatment, medication ~ training, medical treatment

256M, 60W, 39

30

27 24

Epidemiology of Running Injuries

323

Table II. Incidence rates (%) of running injuries based on studies with an incidence period other than 1 year

Reference

Pollock et al. (1977)

Jacobs & Berson (1986) Watson & DiMartino (1987) Bovens et al. (1989)

Effect on training

~

training> 7days

I training Stopped training Stopped training > 2 sessions, I training I training

No. of subjects. sex. average age (years)

Training load

2 groups of male inmates: 87M, 70M 20-35

1. 3 sessions/wk 15, 30 or 45 min 2. 30 min 1, 3, 5 sessions/wk 70 km/wk

2.0,12 & 39

10 h/wk

17.5

24, 35 or 44 km/wk

58,60 & 67 depending on training phase

355M,34 96W, 32 257 boys & girls, 16

58M,35 15W, 34

The incidence rate varies from 24 to 77%. If restricted to studies in which samples of more than 500 subjects were used, the yearly incidence rate varies from 37 to 56%. In table I yearly incidence rates are presented from studies with a data collection period of 1 year. All the studies referred to in this table are cohort studies, either prospective or retrospective. It can be speculated that studies based on a period of data collection not equivalent to a full year do not provide representative yearly incidence rates, since running participation does not necessarily need to be ofthe same intensity during 12 months of the year. For example, Lysholm and Wiklander (1987) registered a significantly higher incidence rate during spring and summer and Walter et al. (1989) calculated that running whole year round was significantly related to a higher incidence rate. In contrast, Yzerman and Van Galen (1989) found that the time of year did not influence the chance of sustaining a running injury. It was found by Van Galen and Diederiks (1990) that 85% of the runners in their study were running the whole year round. In table II incidence rates are presented from 4 studies with a time period of data collection not equivalent to 1 year. Two of these studies are cohort studies with incidence rates varying from 17.5 to 46.5%. The other 2 studies concern trials in which the running load was programmed by the study and,

Injury incidence (%)

1.22,24 & 54

46.5

therefore, as discussed below, largely influenced by the study design. The incidence rates in these 2 studies varied from 0 to 67%. To summarise the data it appears that for the average recreational runner, who is steadily training and who occasionally participates in a long distance run, the overall yearly incidence rate for running injuries varies between 37 and 56%. Depending on the specificity of the group of runners concerned (competitive athletes, average recreational joggers or boys and girls) and in different circumstances these rates may vary (see tables I and II). To compare studies the incidence of sports injuries in relation to exposure in days, hours or sports event can be calculated (Eriksson 1983; Kennedy 1977; Wallace 1988; Zemper 1984). For this purpose, injury incidence is expressed as the number of injuries per 1000 hours of sports participation by many researchers (Backx et al. 1990; Ekstrand et al. 1982; Lysholm & Wiklander 1987; Van Mechelen et al. 1987). Van Galen and Diederiks (1990) found in an overall population survey an injury incidence of 3.6 injuries per 1000 hours of running. For competitive athletes this incidence figure varies from 2.5 to 5.8 injuries per 1000 hours of running depending on the specialisation of the athlete (Lysholm & Wiklander 1987). In the study by Bovens et al. (1989), in which running performance was largely inflicted by the study itself, the incidence

324

Sports Medicine 14 (5) 1992

Table III. Distribution of running injuries according to medical diagnosis; NA = not applicable; figures show the percentage Lysholm & Wiklander (1987) In

= 55)

Watson & DiMartino (1987) In = 41)

Marti et al. (1988) In

= 877)

Tendinitis Inflammation (other

33 4

10

17

15

NA

than tendinitis) Strain

15

15

18 NA

15

5 22

Chondromalacia Periostitis/stress fracture Sprain Other

11 9

17 17

12 14 39

varied from 12.1 to 7.0 injuries per 1000 hours running depending on the weekly running distance. This study also clearly demonstrated that with increasing exposure the incidence per 1000 hours of running decreases, whereas the incidence rate (%) increases as a result of the cumulative effect of increased exposure. This phenomenon was also demonstrated by Marti et al. (1988). If it comes to the risk of running compared with that of other sports it has been shown in Switzerland that running injuries are 2 to 2.5 times less frequent than injuries from all other sports and about 6 times less frequent than ski injuries (Marti et al. 1988). In the Netherlands, according to Van Galen and Diederiks (1990) the overall Dutch sports injury incidence was estimated to be 3.3 injuries per 1000 hours of sports, running/jogging, jointly with squash and motor sports, ranking 14th with an incidence of 3.6 injuries per 1000 hours.

2. Severity of Running Injuries

2.1 Medical Diagnosis

In competitive athletes tendinitis was the most frequent registered injury (Lysholm & Wiklander 1987), in boys and girls, periostitis/stress fracture (Watson & DiMartino 1987), and in a normal jogging population, strain and tendinitis (Marti et al. 1988) [see table III].

The most frequent diagnosis of medically treated running injuries was shown in the results of Clement et al. (1981), who carried out a retrospective clinical survey of 1650 runners, 987 men and 663 women over a 2-year period. In total 1819 injuries had been sustained; the frequency of the 10 most common diagnosis of injury, accounting between them for 69% of all injuries (table IV). All injuries involved the knee, the leg or the foot. Most of the running injuries are associated with overuse (Andrews 1983; Clement et al. 1981; Dressendorfer & Wade 1983; Eggold 1981; Lehman 1984; Mirkin 1975; Stanish 1984). Marti et al. (1988), and Ooijendijk and Van Agt (1990), found in their respective studies about 75% of all injuries to be of an overuse nature, whereas Yzerman and Van Galen (1987) registered 54% overuse injuries. Running injuries tend to reoccur (Walter et al. 1989). Data from the epidemiological studies on the recurrence of injuries, however, are not consistent with figures varying from 21 to 70%. This large range is probably affected by differences in the definition of a recurrent injury and by differences in research methodology. Walter et al. (1989) also noted that with the recurrence of injuries there are major differences between the affected part of the body. Their study had rates of new injuries ranging from 29% for back injuries to 73% for shin injuries. 2.2 Site of Injury In the majority of the epidemiological studies knee injuries account for 25% of the injuries. The feet have between 2 and 22% of injuries, the ankles between 9 and 20%, lower leg between 2 and 30%, shin between 6 and 31 %, upper leg 3 and 18%, back between 3 and 11 %, and hip/pelvis and groin between 2 and 11 % (Blair & Ko~l, 1987; Bovens et al. 1989; Jacobs & Berson 1986; Lysholm & Wiklander 1987; Macera et al. 1989; Marti et al. 1989; Ooijendijk & Van Agt 1990; Walter et al. 1989; Watson & Di Martino 1987; Yzerman & Van Galen 1987). In general, running injuries are located from the knee downwards in about 70 to 80% of all injuries.

Epidemiology of Running Injuries

325

Table IV. Frequency of the 10 most common medically treated injuries sustained by 987 male and 663 female runners (after Clement et al. 1981) Medical diagnosis

Males

Females

Total

%

n

%

n

%

Patellar pain syndrome Tibial stress syndrome Inflammation achilles tendon Fasciitis plantaris Inflammation patellar tendon Iliotibial friction syndrome Metatarsal stress syndrome Tibial stress fracture Tendinitis M. tibialis posterior Tendinitis M. peroneus

24.3 10.7 7.9 5.3 5.6 4.6 3.3 2.4 1.9

262 115 85 57 60 50 36 26 21

27.9 16.8 3.2 3.9 2.8 3.8 3.0 2.8 3.2

206 124 24 28 21 28 22 21 14

25.8 13.2 6.0 4.7 4.5 4.3 3.2 2.6 2.5

468 239 109 85 81 78 58 47 45

2.0

22

1.6

12

1.9

34

Total

68.0

735

69.0

510

68.7

1244

n

NB: Patellar pain syndrome (chondromalacia patellae) is a disorder of the kneecap; fasciitis plantaris is a disorder of the sole of the foot; tibial stress syndrome (shin splints) is caused by strain on the points of attachment of the foot flexors; and tractus iliotibialis

syndrome is pain in the outer side of the knee.

2.3 Consequences of Running Injuries

The severity of sports injuries can also be judged by the impact of the injuries on factors such as sporting time lost, medical treatment and/or absence from work. Running injuries lead to a reduction of training or training cessation in about 30 to 90% of all injuries, about 20 to 70% of all injuries lead to medical consultation or medical treatment and 0 to 5% result in absence from work. These figures must be judged with caution given particularities of the studies involved. Marti et ai. (1988) gave the most extensive description: 44% of 1994 running injuries resulted in cessation of training with an average duration of 4.8 weeks, 31 % led to medical consultation with an average number of 3.8 consultations and 5% to absence from work with an average duration of 10.1 days. Clough et ai. (1989) found significant differences between marathon finishers and dropouts with respect to the length of full cessation of training. 35% of all injured marathon finishers stopped training for 4 weeks or more opposed to 65% of all injured mar-

athon dropouts. H0lmich et ai. (1989) found in a retrospective cohort study (a period of data collection of 1 year), that the percentage of injured runners stopping training as a result of a running injury increased with the distance covered per week: 23% of all runners running 0 to 30 km/week were prevented from running as a result of an running injury, gradually increasing to 53% of all runners running more then 120 km/week. At this stage there are no studies known in which the direct and/or indirect cost of running injuries are calculated.

3. Causes of Musculoskeletal Running Injuries Running injuries are of a diverse nature and vary from metabolic abnormalities such as anaemia, amenorrhoea, hypothermia and hyperthermia to extrinsic hazards such as dog bites and traffic collisions (Powell et al. 1986). However, the discussion on the causes of running injuries is limited to

326

musculoskeletal injuries, the most common running injuries. Almost all of the cited authors have compared injured and noninjured runners by using univariate statistical techniques. However, many of the factors which are identified in this way may interact or be interrelated. For instance in the Yzerman and Van Galen (1987) study after univariate analysis many factors were found to be significantly (p < 0.05) associated with running injuries. A key factor in their analysis was membership of an athletic club which 'pooled' a number of other significant factors such as more malalignment disorders, no participation in other sports, higher weekly running distance, higher running speed, more running on a hard surface, frequent change of running shoes, running more marathons per year, and high motivation to compete. To cope with problems associated with interrelation and interaction of risk factors some authors (Macera et al. 1989; Marti et al. 1988; Walter et al. 1989; Yzerman & Van Galen 1987) have, in addition to univariate analysis, also used multivariate statistical techniques, such as discriminant analysis and stepwise logistical regression (Dixon et al. 1990) giving a clearer insight into the relative importance of risk factors. A summary of factors related to injuries is presented in table V. 3.1 Characteristics of Runners 3.1.1 Age In the studies of Koplan et al. (1982), Blair et al. (1987), Jacobs and Berson (1986), Macera et al. (1989) and Walter et al. (1989), which are comparable with regard to study population (male recreative entrants to road races or members of a fitness club), age proved to be not significantly related with running injuries. Yzerman and Van Galen (1987) showed that there was a significant but low correlation (p < 0.05, r = 0.10) between running injuries and age, runners at a younger age sustaining more injuries then older runners. Younger age also contributed in this study significantly to the prediction of injury after discriminant analysis. However, interpretation of this finding was ques-

Sports Medicine 14 (5) 1992

Table V. Factors related with running injuries: a summary of results from the literature

Factors significantly related with running injuries

Factors not significantly related with running injuries

Factors related with running injuries that are not clear, or are contradicting or based on scanty information

• PreviOUS injury • Lack of running experience

• Age • Gender • Body mass

• Warming up • Stretching exercises

• Running to compete

index • Hill running • Running on hard surface

• Body height • Malalignment • Restricted range of motion

• Excessive weekly running distance

• Participation in other sports • Time of the year

• Running frequency • Intensity of performance

• Stability of running • Time of the day pattern • Shoes • In-shoe orthoses • Running on 1 side of the road

tioned by the authors because younger runners train at a higher speed, also running more kilometres per training than older runners. In a study by Marti et al. (1988), age-stratification showed a significant decrease of running injuries with an increase of age. It was speculated by Marti et al. that this finding could be because of a 'healthy runner effect' by which only those runners remaining free of injuries continue to run, since Marti et al. (1988) also found within the most numerous group of runners (age 33 to 44), after stratification for length of jogging career, a decrease of injuries with an increase of years of running. It was also speculated that this finding could be because of a possible long term adaptive process of the musculoskeletal system. However, after multiple logistic regression age was not confirmed as an independent risk factor. In conclusion, in the majority of studies age was not associated with running injuries or it was speculated that the association was biased by factors such as running experience, weekly running distance or running speed.

327

Epidemiology of Running Injuries

3.1.2 Gender

Powell et al. (1986) concluded that 'gender per se does not seem to be an important risk factor for running injuries'. This appears to be confirmed by the 4 epidemiological studies with enough female subjects to take this factor into account, which found gender to be not related with running injuries (Jacobs & Berson 1986; Lysholm & Wiklander 1987; Walter et al. 1989; Yzerman & Van Galen 1987).

3.1.3 Anthropometries

It may be speculated that taller and/or overweight individuals sustain a greater likelihood of running injuries as a result of greater forces on bones, joints or connective tissue. However, it may be speculated that these greater forces will be balanced, for example, by stronger muscles and larger bone surface areas. Only in 2 studies was the relation between body weight and running injuries evaluated. Yzerman and Van Galen (1987) and Walter et al. (1989) did not find a significant association between bodyweight and running injuries. Body height was only evaluated in 1 study (Walter et al. 1989) showing after age-adjusted univariate logistic regression taller men being at greater risk of injury then shorter men, an effect equivocated when height was adjusted for weight. The association between body mass index (BMI), defined as bodyweight (kg)/body height 2 (m 2 ) and running injuries was evaluated in 5 studies. Koplan et al. (1982) found BMI to be not related to running injuries. A similar finding was reported by Macera et al. (1989) and Walter et al. (1989). In contrast to these findings, Blair et al. (1987) found a significant but low correlation (p < 0.05, r = 0.10) between BMI and running injuries. However, Marti et al. (1988) found runners with a BMI smaller then 19.5 and runners with a BMI greater then 27 to be at greater risk of running injuries, although overweight defined as BMI 26 to 28, did not prove to be a risk factor in this study. It should be noted though that overweight amongst runners is a rare finding. In the study of Marti et al. (1988) only 1.8% of 4358 runners had a BMI greater than 27.

In a study by Helmich et al. (1989) only 8% out of a total of 1426 nonelite marathon runners had a BMI greater than 25. It can from the majority of these studies, with some precaution, be concluded that BMI as such is not related with running injuries. Only I study showed runners with a BMI smaller then 19.5 or greater then 27 to be at significant greater risk, while perhaps taller runners are also at greater risk.

3.1.4 Malalignment

In the literature many alignment defects are associated with overuse running injuries, for example: difference in limb length (Subotnick 1985); femoral anteversion (Stanish 1984); knee anomalies, such as knock knees or bow legs and too large or too small patella (Andrews 1983; Clement et al. 1981; Gudas 1980); foot anomalies, such as varus and valgus of the heel or rear foot, flat feet and high arches (Doxey 1985; McKenzie et al. 1985; Renstrom & Johnson 1985). Powell et al. (1986) noted that no epidemiological study had evaluated the effect of malalignment as a risk factor for running injuries. Since then 2 epidemiological studies have mentioned malalignment to be significantly associated with running injuries. Lysholm & Wiklander (1987) found, in an attempt to analyse the causes of prospectively registered running injuries retrospectively, that in 40% of the cases, malalignment was at least 1 of the factors causing the injury. Malalignment in this study included 'foot insufficiency, lower extremity muscle stiffness, genu varum and high Q-angle'. In their retrospective study by questionnaire, Yzerman and Van Galen (1987) included questions on asymmetry of the lower extremity as assessed during previous medical checkups. After univariate analysis they found a significant correlation (p < 0.05, r = 0.14) between asymmetry and medically treated injuries. This remained significant after multivariate analysis. The epidemiological study of Walter et al. (1988) is the only study in which at baseline a physical assessment of the lower extremity was included. None of the measured variables (femoral neck anteversion, pelvic obliquity, knee and patella alignment and rear foot valgus) were signifi-

328

cantly associated with running injuries. From these scarce and contradicting results the importance of malalignment as a cause of running injuries is stilI not made clear. However, as stated by Powell et al. (l986), 'the hypothesis that structural abnormalities are a risk factor for running injuries is too reasonable to deny' and 'careful, abnormality specific studies should be a top priority for future research'.

3.1.5 Muscular Imbalance, Restricted Range of Motion Running strengthens the muscles at the back of the thigh and leg (the hamstrings and calf muscles) relatively more than those at the front (Clement et al. 1984; Mirkin 1975; Subotnick 1985), and the resulting imbalance supposedly facilitates the occurrence of various types of overuse injuries. A weak vastus medialis is thought to be a causal factor in damage to the patellar cartilage (Stanish 1984). It is also speculated that overuse injuries may result from a lack of flexibility, the physiological shortening of the muscle through tiredness, insufficient muscular strength and endurance, a strength imbalance between the left and right hamstrings, an imbalance between the hamstrings and the quadriceps, asymmetric contraction of the hamstrings and too early a resumption of sporting activity following an injury (Van Mechelen et al. 1987). No epidemiological studies mention this factor. More research is needed. 3.1.6 Running Experience Running experience is thought to be a factor associated with running injuries, the inexperienced runner being more susceptible to running injuries (Powell et al. 1986). .This factor did not prove to be significantly associated with running injuries in the studies of Koplan et al. (1982), Blair et al. (1987), Yzerman & Van Galen (1987) and Walter et al. (1989). Marti et al. (1988) found in an age-specific analysis among men of similar age a positive association between years of jogging and decreased incidence of running injuries. In the study of Macera et al. (1989) after univariate analysis it was found

Sports Medicine 14 (5) 1992

for men that running regularly for less than 3 years was significantly associated with running injuries of the lower extremity. This association remained after multivariate logistic regression: runners with less than 3 years of running experience had an odds ratio of 2.2 [95% confidence interval (CI): 1.5-3.3]. These findings suggest that for men, running experience, resulting in a better adaptation to running on the biomechanical and/or tissue level, may lead to fewer injuries, although bias as a result of a 'healthy runner effect' should be taken into consideration (Marti et al. 1988).

3.1.7 Previous Injury According to Powell et al. (1986) persons with a previous injury may be more likely to be injured again because: the original cause may remain; the repaired tissue may function less well or be less protective then the original tissue; or the injury may not have healed completely. In 3 epidemiological studies, previous injury defined as 'a positive history of running injury in the 12 months prior to the start of the study', was associated with running injuries. Marti et al. (1988) found in univariate analysis, after adjustment for differences in weekly running distance a 65% increased risk on injury for runners with a history of previous injury. This remained after multivariate analysis. Macera et al. (1989) found for men a history of previous injury to contribute significantly to injury after univariate age-adjusted logistic regression (odds ratio of 2.7; 95% CI: 1.9-3.9). After multivariate analysis the odds ratio remained the same. Walter et al. (1989) found after univariate age-adjusted logistic regression for men a significant odds ratio for previous injury of 1.7 (95% CI: 1.2-3.5) and for women of 2.4 (95% CI: 1.3-4.1). After multiple logistic regression in which age and sex were forced into the model, previous injury proved to be I of the 2 strongest predictors for running injury with an odds ratio of 1.5. From these studies it can be concluded that previous injury as defined above is an important independent risk factor for running injuries, although the mechanisms still needs clarification.

Epidemiology of Running Injuries

3.1.8 Participation in Other Sports It is speculated that runners participating in

other sports sustain less running injuries then runners who do not do so. Reasons for this phenomenon may be that runners who run more miles per week have less time for participating in other sports, also that runners who spend more time in other sports use different muscle groups and in this way decrease their risk on overuse injuries associated with running (Jacobs & Berson 1986). Jacobs and Berson (1986) found injured runners participated significantly less in other sports in comparison with noninjured runners. However, Yzerman and Van Galen (1987) found no relationship between participation in other sports at least once a week and running injuries. Also Marti et al. (1988) found that regular practice of other sports was not associated with reduced running injury risk. Finally, Walter et al. (1989) also showed that participation in other sports was not associated with the risk of running injury. Similar risks were found in runners who reported to be active in a large variety of forms of activity. Only ice skating and dancing showed a significant relation to running injuries with about one-third less risk of injury. In this study also the risk of injury was studied according to the usual proportion of time spent sitting, standing or walking during the day; no significant relationship could be established. From these findings it can be concluded that participation in other sports is not a independent risk factor. However, this point needs further study. 3.1.9 Psychological and Behavioural Factors Notwithstanding the presence of any evidence in relation to running, Powell et al. (1986) presumed psychological factors to be related to running injuries. Yzerman and Van Galen (1987) showed that the injury incidence was significantly related to a motivation score. The higher the motivation score the larger the incidence for medically treated injuries. They speculated that more motivated runners ignore first signs of injury more than less motivated runners, therefore, having a higher injury incidence. After multivariate analysis motivation still contributed significantly to the risk of

329

sustaining a running injury. Both Marti et al. (1988) and Walter et al. (1989) found competitive running (e.g. runners who stated that they run predominantly to compete), in contrast to running for fitness or recreation, to be significantly associated with running injuries. In the Marti et aI. (1988) study this association proved to be independent after multivariate analysis and Walter et al. (1989) argued that the higher injury risk of competitive runners was probably because of the higher training mileage. It can be concluded that psychological and behavioural factors seem important and need further future investigation. 3.2 Characteristics of Running Practice 3.2.1 Weekly Running Distance, Running Frequency and Weekly Running Time Almost all of the cited authors state that weekly running distance is the most important risk factor for running injuries. In a randomised trial Pollock et al. (1977) found weekly running frequency and duration of running to be related with running injuries. Both increasing duration and increasing frequency produced more injuries. Koplan et al. (1982) found an almost linear relationship between increasing weekly distance and running injury incidence for both men and women. Jacobs and Berson (1986) showed weekly distance to be significantly associated with running injury. The same was confirmed in epidemiological studies of Blair et al. (1987), Yzerman and Van Galen (1987), Marti et al. (1989), Macera et al. (1989), Walter et al. (1989) and Bovens et al. (1989). Lysholm and Wiklander (1987) found running injuries to be significantly associated with monthly running distance. Only I epidemiological study did not confirm the relation between exposure time and injuries (Watson & DiMartino 1987). In this study, exposure time was defined as 'time spent training and competing on a weekly basis'. No difference in exposure time was found between injured and no.1injured athletes. However, the population in this study was different (see table I) from the above mentioned studies in terms of age and participation in athletic events.

330

However, despite the absolute risk of running injury increasing with increasing weekly distance, the risk per exposure time decreases as was shown by Marti et al. (1988), Walter et al. (1989) and Bovens et al. (1989). For prevention it is possible to speculate about the effect of dividing total weekly distance over several shorter sessions if running injuries are related to weekly distance. In some studies a significant relationship between running frequency (Jacobs & Berson 1986; Macera et al. 1989; Walter et al. 1989) and running injuries was found, although this relation was not confirmed by Marti et al. (1988). Marti et al. (1988) examined a subgroup running the same weekly distance in 2, 3 or 4 weekly sessions. No significant differences appeared in the incidence of running related injuries. In the study of Jacobs and Berson (1986) duration of running was not associated with increased running injury incidence. It should be considered that these variables are strongly interrelated. In conclusion, it can be stated that weekly running distance is a strong determinant of running injuries. The role of running frequency and weekly running time remains unclear. 3.2.2 Training Speed and Other Indicators of Intensity of Running Performance Training speed was not associated with increased risk on running injuries in the studies of Koplan et al. (1982), Yzerman and Van Galen (1987) and Walter et al. (1989). In the study of Jacobs and Berson (1986), however, training speed was significantly associated with an increased risk of running injuries, as in a subpopulation of younger runners in the study of Yzerman and Van Galen (1987). In addition to training speed, intensity of running performance can, for instance, also be operationalised by participating in marathon running, fastest time run on a certain distance, running whole year round, and running at least once a week 8km or more. Watson and DiMartino (1987) found performance level based on a percentile ranking of the best season performance to be significantly associated with injury. Yzerman and Van Galen (1987) found

Sports Medicine 14 (5) 1992

marathon running and the best time on a half marathon to be significantly associated with an increased injury risk. Marti et al. (1988) showed the best time on a 16km run to be significantly associated with running injury. Macera et al. (1989) found running of at least 1 marathon a year to be significantly associated with injury risk. Walter et al. (1989) found running whole year round and running at least once a week 8km or more to be significantly associated with running injury. Frequency of race competition in this study was not associated with running injury. From these findings the relation between running speed and/or intensity of running performance and injury risk remains unclear. 3.2.3 Stability of Running Habits A sudden increase in weekly running distance or other training habits has been associated with a cause of running injuries, because of lacking capacity of tissue to adapt to this kind of sudden change (Powell et al. 1986). Training errors such as running too often, too fast or too long are major causes of injury in both beginners (Franklin et al. 1979; Kowal 1980) and experienced runners. Any sudden increase in the weekly distance (Koplan 1982) or any change to a specific form of training (such as interval training or hill training) without a gradual buildup is regarded as a training error (Andrews 1983; Clement et al. 1984). Lysholm and Wiklander (1986) found training errors (e.g. sudden change in training routine) to be associated with running injury in about 60% of all injuries. Yzerman and Van Galen (1987) found a low but significant correlation (p < 0.05, r = 0.08) between irregular training and running injuries. However, in this study also a significantly higher injury incidence was found in runners training according to a training scheme set out by a trainer (53%) compared with runners running without such a scheme (35%). In the study of Walter et al. (1988) no association was found between 'the use of hard training days' and running injuries. Running sprints or intervals was not related to running injuries in the study of Jacobs & Berson (1986). In the same study 33% of the injured run-

Epidemiology of Running Injuries

ners had changed either their training technique or running shoes prior to their injury. 3.2.4 Warm-Up, Stretching Exercises and Cooling Down Lack or improper use of warm-up, stretching exercises and cooling down is thought to be a risk factor with regard to running injuries (Van Mechelen et al. 1987). Jacobs and Berson (1986), as well as Yzerman and Van Galen (1987), found injured runners to have stretched significantly more before running than noninjured runners. Jacobs and Berson (1986) claimed that certain stretching exercises such as the hurdler stretch can lead to injury of the medial collateral ligament and to the medial meniscus. Both studies put forward that runners who are injured stretch because of their injury. Walter et al. (1989) found runners who occasionally stretch to be significantly at risk of injuries in contrast to runners who 'always, usually or never stretch' leaving this finding unexplained. Blair et al. (1987) found frequency of stretching not to be associated with running injuries. Macera et al. (1989) found stretching before running to be not associated with running injury. Walter et al. (1989) found that runners who say they 'never' warm up run a significantly smaller risk of running injury than runners who say they 'always, usually or sometimes' warm up. In the same study regular use of cool-down was not related to running injuries at all. In conclusion, there seems to be a negative, rather than a positive relation between these preventive measures and the risk off running injuries, although the findings are inconclusive. Further studies are required.

3.3 Characteristics of the Running Environment 3.3.1 Shoes Running is a repetitive movement comprising 2 phases, foot contact with the ground (support phase) and when feet are off the ground (airborne phase). During the ground contact phase the weight of the body is absorbed by the foot and then lifted

331

as the next stride begins. Two types of landing are distinguished, namely on the heel ('rear foot strikers'; 80% of all runners) and on the front part of the foot (mid- and forefoot strikers; 20% of all runners); types of landing intermediate between the 2 may also occur, depending on speed. With respect to type of landing Jacobs and Berson (1986) did not find any difference in injury incidence in relation to the part of the foot first contacting the ground. The forces involved in landing may be as great as 2 to 5 times bodyweight (Cook et al. 1990; McKenzie et al. 1985; Subotnick 1985). A runner with a stride of l.5m makes contact with the ground about 670 times per kilometre, and if we assume a body weight of 70kg and an average force on landing of 2.5 times bodyweight this implies that each leg is subject to a force of around 60 tonnes/ km. This shock is absorbed by the thigh and leg muscles and by the movement of the foot, with pronation important. It seems evident that a training shoe providing cushioning, support and stability can play an important role in shock absorption, and thereby injury prevention (Cook et al. 1990; Robbins & Gouw 1990), although there are also data suggesting that modern running shoes protect poorly from running injuries and may cause chronic overloading (Robbins & Gouw 1990). Yzerman and Van Galen (1987) found a significant correlation (p < 0.05, r = 0.10) between the regular change of shoes and medically treated running injuries. A possible explanation for this phenomenon is the fact that injured runners try to solve their problem by changing shoes. Marti et al. (1988) found in their study no difference in injury incidence between runners preferring I of the 3 most popular running shoes. Runners having no preference for any brand of shoes sustained significantly fewer injuries than runners who had a preference for a specific brand. From this finding the conclusion was drawn that no particular shoe has a preventive advantage over other brand names. Runners wearing inexpensive shoes « approximately $US39) did not sustain more injuries, whereas runners wearing expensive shoes (dearer than about $US90) sustained significantly more in-

332

juries. A likely explanation put forward by Marti et al. (1988) is that injury-prone runners or high mileage runners choose to wear expensive shoes. It is said that the use of orthotic devices is capable of correcting minimal alignment abnormalities, in this way creating normal running conditions. Marti et al. (1988) found the use of orthotics to be significantly related with both a positive history of previous injury and with running-related injuries during the period of study. Selection bias with injury-prone runners favouring orthotics was considered as an explanation. Walter et al. (1989) found owning 2 pairs of shoes to be significantly related with a 50% increase in injury risk. However, in this study shoe characteristics (presence of varus wedge, wear sole, wear pattern and personal shoe repair) were not related with injury risk. The effect of owning more shoes reflected greater levels of training of such runners. Although the role of running shoes in running injury prevention remains unclear, the importance of shock absorbency by running shoes does not appear to be denied by the cited epidemiological studies. In addition, Cook et al. (1990) pointed out that the shock absorbing qualities of wet running shoes is reduced and that all running shoes lose between 30 and 50% of their shock-absorbing characteristics after about 400km of running. 3.3.2 Running Surfaces and Hill Running Running on hard surfaces increases mechanical shock and may overload joints and tendons (Clement & Taunton 1981), while soft surfaces quickly tires the muscles and it is thought that it may increase injury (Gudas 1980). Surface irregularities for example, potholes, the roots of trees, pavement edges may cause acute injury to the ankle and knee. Running on roads, not only through the hardness of the surface but also through its camber, which places a different loading on the legs, may also increase overuse injuries of various types, e.g. short leg syndrome (Gudas 1980). Jacobs and Berson (1986) did not find a correlation between running surface and running injury, but this finding was probably biased by the

Sports Medicine 14 (5) 1992

majority of the runners in their study running on concrete and asphalt. The same was found by Yzerman and Van Galen (1987), who added bias as a result of a 'healthy runner effect' (all runners run on hard surfaces and only those who remain not injured are still running on hard surfaces) as a possible explanation. Marti et al. (1988) was not able to prove an advantage or disadvantage from the usual running surface (predominantly hard, predominantly natural or combined). Walter et al. (1989) found no significant relation between injuries and frequency of running on asphalt, concrete, grass or dirt. Macera et al. (1989) found, after both univariate and multivariate analysis, in women a significant difference in injury risk for running on a concrete surface (odds ratio 5.6; 95% CI: 1.1-29.3). This relation was not found in men. Bovens et al. (1989) found injuries of the lower leg and achilles tendon to be significantly more localised on the left hand side. This was explained by the unequal load ofthe leg caused by the surface of the streets, which is a little convex in order drain away the rain water. From these findings it appears that in men, different running surfaces do not influence running injury risk, although it should be kept in mind that bias probably plays an important role. Hill running is associated with increased injury risk (Clement et al. 1981). Jacobs and Berson (1986), Macera et al. (1989) and Walter et al. (1989) did not find any relation between hill running and increased injury risk. 3.3.3 Time of Day and Season It has been suggested that persons who run in

the morning are more likely to be injured than those who run at other times of the day (Powell et al. 1986), although this finding was not confirmed by Blair et al. (1987), Jacobs and Berson (1986), Macera et al. (1989) and Walter et al. (1989), who all found no association between time of day and increased injury risk. Powell et al. (1986) suggested that because of climatic circumstances, such as snow, roads may be slippery causing falls leading to more injuries during winter and fall. In only 1 study was climate

Epidemiology of Running Injuries

taken into consideration. Yzerman and Van Galen (1987) found no relation between running injuries and time of year, although agreeing that this finding could be biased by unequal exposure whole year round. Walter et al. (1989) found running whole year round to significantly enlarge the risk on running injury. This may either reflect climatic influence or differences in exposure.

4. Prevention Through Health Education From the available data it seems clear that previous injury, lack of running experience, running to compete and excessive weekly running distance are significantly associated with running injuries. The association between running injuries and factors such as warm-up and stretching exercises, body height, malalignment, muscular imbalance, restricted range of motion, running frequency, intensity of running performance, stability of running pattern, shoes and orthotic devices and running on I side of the road remains unclear or is backed by contradicting or scarce research findings. Significantly not associated with running injuries seem age, gender, body mass index, running hills, running on hard surfaces, participation in other sports, time of the year and time of the day. These data are summarised in table V. The fact that information on some factors is contradicting or scarce and that other factors do not seem to be related to running injuries does not necessarily mean that these factors are of no importance. As stated by Walter et al. (1989), this merely reflects that in epidemiological research the range of effects of these factors is probably not important within the range of the majority of runners, but they can undoubtedly be causative for some individual injuries. Measures aimed at preventing running injuries must be geared to the aetiological factors involved (Clement & Taunton 1981). However, the causes of running injuries are so multifactorial and diverse that any specific single measure proposed would probably be of help to only a small minority of runners (Marti et al. 1988). Perhaps a multifactorial concept regarding the aetiology and preven-

333

tion of sports injuries as outlined by Van Mechelen et al. (1992) may be helpful in this respect. A common means of prevention is by providing information by health education. Health education is only effective if it is put forward as a planned strategy (Kok & Bouter 1990). Kok and Bouter (1990) argued that such a planned strategy should be aimed at a favourable modification of the determinants of the correct behaviour. The implications of the risk factors discussed in this chapter speak, in terms of preventive measures, for themselves and will not be further discussed. For those factors undoubtedly significantly related with running injuries some remarks with regard to preventive measures will be made, but only in the perspective of the modification of (un-)healthful behaviour in relation to running. 4.1 Complete Rehabilitation and Early Recognition of Symptoms of Overuse Dealing with the risk factor 'previous injury', complete rehabilitation and early recognition of symptoms of overuse injuries seem very important and can be regarded as determinants of correct behaviour with respect to running injuries. Without going into the precise nature of a rehabilitation programme, some authors state that it can prevent the recurrence of injuries to the musculoskeletal system (Davies 1981; Ekstrand 1982; Harvey 1982; Hayes 1974; Lysens et al. 1984; Proctor 1980; Runyan 1983; Stulberg 1980; Thompson et al. 1982; Watson 1984). Complete rehabilitation should prevent an injured sportsman restarting the sporting activities too soon. A rehabilitation programme cannot be regarded as having been completed until (Ekstrand 1982; Proctor 1980): the sportsman is free from pain; muscle strength has returned to about the preinjury level; and articulatory mobility has recovered to preinjury level. In cases of articulatory instability the purpose of rehabilitation is to strengthen the muscles around the unstable joint, compensating for the instability and thus reducing the risk of repeated injury to a ligament or tendon (Runyan 1983). We are not aware of any results of

334

research into the effects of a rehabilitation programme with respect to running, although it has been demonstrated that a whole package of measures including 'controlled rehabilitation' can reduce the number of sports injuries in soccer (Ekstrand 1982). With regard to early recognition of overuse injuries a runner should be taught to listen and to respect 'the language of his body' and reduce or temporarily stop, rather than to continue or increase, running when suffering from pain or stiffness of joints and tendons as a result of running. 4.2 Training Guidelines Lack of running experience and running to compete are risk factors which are hard to control in terms of prevention other than through health education. With regard to lack of running experience some general training guidelines were formulated by Van Mechelen et al. (1987): training should be built up gradually; running speeds should be such that the runner can continue to speak without shortness of breath (Kemper 1985); the untrained should start their training gradually, e.g. on alternate days; in group training each individual should have his or her own training programme (this is helped by a division into groups of more or less equal performance). Excessive weekly distance is also a risk factor for running injuries. This factor can be dealt with by reducing running distance itself However, this does not seem to be an acceptable measure for joggers who are determined to run. At the level of the individual runner the question 'How much is too much?' can only be answered by trial and error. In principle a training consisting of running shorter distances with higher intensities (higher running speed) has an equal or even better effect on the aerobic system. Nevertheless it should be kept in mind that increasing running speed may increase injury incidence, although the relationship between running speed and/or intensity of running performance and injury risk remains unclear. A health education programme should be aimed at a change of running behaviour by reducing excessive run-

Sports Medicine 14 (5) 1992

ning distances and/or by reducing participation in long distance runs such as half and full marathons.

Acknowledgements This study was partly supported with a research grant of the Dutch Ministry of Welfare, Health and Cultural Affairs (grant number 87-33) as a contribution to the Council of Europe research project 'Sport for All: sports injuries and their prevention'.

References Andrews JR. Overuse syndromes of the lower extremity. Clinics in Sports Medicine 2: 137-148, 1983 Backx FJG, Inklaar H, Koornneef M, Van Mechelen W. Draft FIMS position statement on the prevention of sports injuries. Geneeskunde en Sport: special issue, 1990 Blair SE, Kohl HW, Goodyear NN. Rates and risks for running and exercise injuries: studies in three populations. Research Quarterly for Exercise and Sport 58: 221-228, 1987 Bar-Or O. Pediatric sports medicine for the practitioner, SpringerVerlag, New York, 1983 Bovens AMP, Janssen GME, Vermeer HGW, Hoeberigs JH, Janssen MPE, et al. Occurrence of running injuries in adults following a supervised training program. International Journal of Sports Medicine 10: S I 86-S 190, 1989 Clement DB, Taunton JE, Smart GW, McNicol KL. A survey of overuse running injuries. Physician and Sportsmedicine 9: 4758, 1981 Clement DB, Taunton JE, Smart GW. Achilles tendinitis and peritendinitis: etiology and treatment. American Journal of Sports Medicine 12: 179-184, 1984 Clement DB, Taunton JE. A guide to the prevention of running injuries. Australian Family Physician 10: 156-164, 1981 Clough PJ, Dutch S, Maugham RJ, Shepherd J. Pre-race dropout in marathon runners: reasons for withdrawal and future plans. British Journal of Sports Medicine 21: 148-149, 1987 Clough PJ, Shepherd J, Maugham RJ. Marathon finishers and pre-race drop-outs; British Journal of Sports Medicine 23: 97101, 1989 Cook SD, Brinker MR, Mahlon P. Running shoes: their relation to running injuries. Sports Medicine 10: 1-8, 1990 Davies JE. Injuries and physical stress. Journal of Biosocial Science. (Suppl. 7): 151-157, 1981 Dixon WJ, Brown MB, Engelman L, Jennrich Rl. BMDP statistical software manual, Vol. 2, University of California Press, Berkely, 1990 Doxey GE. Management of the metatarsalgia with foot orthotics. Journal of Orthopaedic and Sports Physical Therapy 6: 324333, 1985 Dressendorfer RH, Wade ChE. The muscular overuse syndrome in long distance runners. Physician and Sportsmedicine II (II): 116-130, 1983 Eggold JF. Orthotics in the prevention of runners' overuse injuries. Physician and Sportsmedicine 9 (3): 125-131, 1981 Eichner ER. Exercise and heart disease: epidemiology of the 'Exercise Hypothesis'. American Journal of Medicine 75: 10081023, 1983 Ekstrand J. Soccer injuries and their prevention, Linkoping University Medical Dissertations no. 130, Linkoping, 1982 Eriksson EKG. Sports accidents. Paper presented at a symposium on accidents in Europe, Newcastle upon Tyne, September, 1983

Epidemiology of Running Injuries

Franklin BA, Lussier L, Buskirk E. Injury rates in women joggers. Physician and Sportsmedicine 7: 105-112, 1979 Gudas CJ. Patterns of lower-extremity injury in 224 runners. Comprehensive Therapy 6: 50-59, 1980 Hartung GH. Jogging - the potential for prevention of heart disease. Comprehensive Therapy 6: 28-32, 1980 Harvey J. The preparticipation examination of the child athlete. Clinics in Sports Medicine I: 353-369, 1982 Hayes D. Risk factors in sport. Human Factors 16: 454-458, 1974 Hess GP, Capiello WL, Poole RM, Hunter Se. Prevention and treatment of overuse tendon injuries. Sports Medicine 8: 371384, 1989 Helmich P, Christensen SW, Darre E, Johnsen F, Hartog T. Nonelite marathon runners: health, training and injuries. British Journal of Sports Medicine 23: 177-178,1989 Hughson R. Canadian association of sports sciences position healthful on protection from exertional heat injuries. Canadian Journal of Applied Sports Sciences 6: 99-100, 1981 Jacobs SJ, Berson BL. Injuries to runners: a study of entrants to a 10 000 meter race. American Journal of Sports Medicine 14: 151-155, 1986 Kemper HCG. Gezondheid in beweging, De Vrieseborch, Haarlem, 1985 Kennedy MC, Vanderfield GK, Kennedy JR.sport: assessing the risk. Medical Journal of Australia 2: 253-254, 1977 Kok G, Bouter LM. On the importance of planned health education: prevention of ski injury as an example. American Journal of Sports Medicine 18: 600-605, 1990 Koplan JP, Powell KE, Sikes RK, Shirley RW, Campbell Ge. An epidemiological study of the benefits and risks of running. Journal of the American Medical Association 248: 3118-3121, 1982 Kowal OM. Nature and causes of injuries in women reSUlting from an endurance training program. American Journal of Sports Medicine 8: 265-269, 1980 Lehman WL. Overuse syndromes in runners. American Family Physician 29: 157-61, 1984 Lysens R, Lefevre J, Ostyn M. The predictability of sports injuries: a preliminary report. International Journal of Sports Medicine 5 (Suppl.): 153-155, 1984 Lysholm J, Wiklander J. Injuries in runners. American Journal of Sports Medicine 15: 168-171, 1987 Macera CA, Pate RR, Powell KE, Jackson KL, Kendrick JS, et al. Predicting lower-extremity injuries among habitual runners; Archives of Internal Medicine 149: 2565-2568, 1989 Manders TH, Kropman J. Sportbeoefening drempels en stimulansen. Instituut voor toegepaste sociologie, Nijmegen, 1982 Manders TH, Kropman J. Sport: ontwikkelingen en kosten. Instituut voor toegepaste sociologe wetenschappen, Nijmegen, 1987 Marti B, Vader JP, Minder CE, Abelin T. On the epidemiology of running injuries. American Journal of Sports Medicine 16: 285-294, 1988 McKenzie DC, Clement DB, Taunton JE. Running shoes, orthotics and injuries. Sports Medicine 2: 334-347, 1985 Mirkin G. The prevention and treatment of running injuries. Journal of the American Podiatric Association 66: 880-884, 1976 Ooyendijk WTM, Van Agt L. Preventie van hardloopblessures. Geneeskunde en Sport 23: 146-151, 1990 Pollock ML, Gettman LR, Milesis LA, Bak MD, Durstine L, et al. Effects of frequency and duration of training on attrition

335

and incidence of injury. Medicine and Science in Sport 9: 3136, 1977 Powell KE, Kohl HW, Caspersen CJ, Blair SA. An epidemiological perspective on the causes of running injuries. Physician and Sportsmedicine 14 (6): 100-114, 1986 Powell KE, Paffenbarger RS. Workshop on epidemiologic and public health aspects of physical activity and exercise, a summary. Public Health Reports 100: 118-126, 1985 Proctor -A. The medical responsibilities of the athletic trainer. North Carolina Medical Journal 41: 519-520, 1980 Renstrom P, Johnson RJ. Overuse injuries in sports: a review. Sports Medicine 2: 316-333, 1985 Robbins SE, Gouw GJ. Athletic footwear and chronic overloading: a brief review. Sports Medicine 9: 76-85, 1990 Runyan DK. The preparticipation examination of the young athlete: defining the essentials. Clinical Pediatrics 22: 674-679, 1983 Safran MR, Seaber AV, Garrett WE. Warm-up and muscular injury prevention: an update. Sports Medicine 8: 239-249, 1989 Stanish WD. Overuse injuries in athletes. Medicine and Science in Sports and Exercise 16: 1-7, 1984 Stulberg SD. Sports injuries and arthritis. Comprehensive Therapy 6: 8-11, 1980 Subotnick SI. The biomechanics of running: implications for the prevention of foot injuries. Sports Medicine 2: 144-153, 1985 Thompson JR, Andrish IT, Bergfeld JA. A prospective study of a preparticipation sports examinations of 2670 young athletes: methods and results. Cleveland Clinical Quarterly 49: 225-233, 1982 Van Galen W, Diederiks J. Sport blessures breed uitgemeten, Uitg. De Vrieseborch, Haarlem, 1990 van Mechelen W, Hlobil H, Kemper HCG. How can sports injuries be prevented? NISGZ publication no. 25E, Papendal, The Netherlands, 1987 van Mechelen W, Hlobil H, Kemper HCG. Incidence, severity, aetiology and prevention of sports injuries: a review of concepts. Sports Medicine 14: 82-89, 1992 Wallace RB. Application of epidemiological principles to sports injury research. American Journal of Sports Medicine 16 (Suppl. I): 22-25, 1988 Walter SO, Hart LE, Mcintosh JM, Sutton JR. The Ontario cohort study of running-related injuries. Archives of Internal Medicine 149: 2561-2564, 1989 Watson AWS. Sports injuries during one academic year in 6799 Irish school children. American Journal of Sports Medicine 12: 65-71, 1984 Watson MD, DiMartino PP. Incidence of injuries in high school track and field athletes and its relation to performance ability. American Journal of Sports Medicine 15: 251-254, 1987 Williams AF. When motor vehicles hit joggers: an analysis of 60 cases. Public Health Reports 96: 448-451, 1981 Yzerman JC, Van Galen WCe. Blessures bij lange afstandlopers. Report, KNAU, 1987 Zemper ED. NCAA injury surveillance system: initial results. Paper presented at Olymp. Sc. Congress, Los Angeles, 1984

Correspondence and reprints: Dr Willem van Meche/en, Department of Health Science, Faculty of Human Movement Sciences, Vrije Universiteit en University of Amsterdam, Meibergdreef 15, 1105 AZ Amsterdam, The Netherlands.