Article Text

Download PDFPDF

Fifth metatarsal fractures among male professional footballers: a potential career-ending disease
  1. Jan Ekstrand1,
  2. C Niek van Dijk2
  1. 1Department of Medical and Health Sciences, Football Research Group, Linköping University, Linköping, Sweden
  2. 2Department of Orthopaedic Surgery, The Academic Centre, University of Amsterdam, Amsterdam, The Netherlands
  1. Correspondence to Professor Jan Ekstrand, Department of Medical and Health Sciences, Football Research Group, Linköping University, Solstigen 3, Linköping S-589 43, Sweden; jan.ekstrand{at}


Background There is little information about Metatarsal Five (MT-5) fractures for specific sports.

Objective To study the occurrence, the imaging characteristics, the lay-off times and healing problems of MT-5 fractures among male footballers.

Methods Sixty-four European elite teams were monitored from 2001 to 2012. x-Rays were collected and classified by the Torg criteria.

Results Of 13 754 injuries, 0.5% (67) proved to be MT-5 fractures. Their incidence was 0.04 injuries/1000 h of exposure. A team of 25 players might thus expect an MT-5 fracture every fifth season. Of these fractures, 67% (38) were primary and 33% were refractures. One of the 38 primary fractures was an avulsion of the tuberosity; all the others (97%) located towards the base. In total, 32% of the players with MT-5 fracture were younger than 21 years, 40% of the fractures occurred during the preseason and 45% of the players had prodromal symptoms. In total, 54% of the initial x-rays were classified as Torg type II (stress fractures), and 46% were classified as Torg type I (acute type). After surgical treatment the fractures healed faster, compared with conservative treatment (75% vs 33%, p<0.05). There was no significant difference in lay-off days between players that had been operated, and those that had not (80 vs 74 days, p=0.67).

Conclusions The majority of MT-5 fractures are stress fractures, and mainly occur among young players. There are frequent healing problems, which might be explained by the stress nature of the injury. After surgery there are less healing problems, compared with those in conservative treatment.

  • Athletics
  • Epidemiology
  • Foot Injuries
  • Soccer
  • Stress Fracture

Statistics from

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.


Several studies have shown that proximal Metatarsal Five (MT-5) fractures are difficult to treat, and prone to healing problems such as delayed unions or refractures.1–5 Dameron1 was the first to suggest, in 1975, that athletes should be operated early, to expedite bone-union. Since then, reports have shown a high rate of healing problems after conservative treatment.2–4

Recently, Kerkhoffs et al6 in an overview of the current literature, suggest that there is a consensus—that operating on fractures of the base of MT-5 produces a faster recovery.

Unfortunately majority of MT-5 fracture reports derive from a heterogeneous patient population with a complete mixture of sex, age and sports.3 ,5 There is limited empirical information about MT-5 fractures in specific sports.7 Low et al8 studied MT-5 fractures in American football and reported 11% of healing problems after surgery and 20% after conservative treatment. But we cannot find any specific studies of MT-5 fractures among footballers (soccer players).

The main aim of this study was to determine the incidence, the location and lay-off times of MT-5 fractures among a homogeneous group of male professional football players.

A further aim was to evaluate the radiological characteristics of these fractures and to correlate these findings with the response to treatment. We also analysed healing and complication rates after surgical versus non-surgical treatment. Our enthymeme was that MT-5 fractures in football are rare injuries with a high rate of healing problems.

Material and methods

Study population

Sixty-four male professional football teams (3487 players and 7525 player seasons) from the top divisions of 15 countries in Europe were monitored over a number of seasons (from 1 to 11 seasons), between July 2001 and June 2012. Our full methodology and the validation is reported elsewhere.9 Teams were followed during the entire football season, and also their preseason.

Study procedure and validity

The procedure followed the guidelines established by FIFA's (Fédération Internationale de Football Associations) consensus document10 and by UEFA (Union of European Football Associations).9 To ensure that information was reliably recorded, the teams were provided with a study-manual which contained definitions and described the correct procedures for recording data. To avoid language problems, the manual and the study forms were translated from English into French, Italian, Spanish, German and Russian. In addition, all reports were checked by us every month, and feedback sent to the teams, so as to correct any missing or unclear data.

‘Baseline’ data for the players were collected annually, at the beginning of each season. Each player's training sessions and matches (that is to say, his extent of exposure, in minutes) was recorded by the club's contact person on a Standard Exposure Form. This was sent to us on a monthly basis. This differentiated between the first-team and second-team exposures, as well as national-team exposure, if any. The team medical staff recorded injuries on a Standard Injury Form which was also sent to us every month. This injury form provided information about the diagnosis, the nature and the circumstances of the injury. Everything was recorded as an ‘injury’ which resulted in that the player being unable to train fully, or to play matches—videlicet, time-loss injuries—and said the player was considered injured until the team's medical staff allowed full training and declared him available for selection. Injuries were categorised under four degrees of severity, based on the number of days in absentia. All injuries were monitored until the final day of rehabilitation.

Our definitions are shown in table 1. A metatarsal fracture was recorded as such after clinical examination by the team's medical staff, supported by with radiological evidence. After teams had reported an MT-5 fracture using the Standard Injury Forms, they were then asked to complete a special MT-5 Fracture Form with information about preinjury status (previous surgery, or previous foot injury), type of radiology, fracture type (traumatic or stress, in the opinion of the team's doctors), fracture location, fracture type according to Torg et al,11 treatment and rehabilitation. The team doctors were also asked to supply initial x-rays or other radiological examinations. An independent radiologist specialised in extremity imaging but unaware of clinical findings and lay-off data classified all images into Torg types (see table 1). The end points of treatment were: healing (return to full team-training or matches), refracture or non-union (the decision being made by team doctors and radiologists).

Table 1

Operational definitions

Inclusionary and exclusionary criteria

All contracted players in the first teams were invited to participate in the study and all accepted. Players who had left the team during the season, for example, because of transfer, were only included while with the team. Players were excluded, if they had had MT-5 fractures within the previous 10 years. Seven cases were excluded, because although reported as MT-5 fractures by team doctors, there were no corroborating fracture-lines when imaged, or because images were lacking (figure 1).

Figure 1

Flow chart of fifth metatarsal fractures.


Association between categorical variables was measured with Pearson's χ2 test or Fisher's exact test. Incidence was calculated as the number of injuries per 1000 h of exposure. The significance level was set at p<0.05. All statistical analyses were made in IBM SPSS Statistics V.19.0. The study design underwent an ethical review and was approved by the UEFA Football Development Division and the Medical Committee.


Injury incidence

A total of 67 MT-5 fractures were registered in 53 players (see figure 1). MT-5 fractures represented 0.5% of all injuries. The incidence of MT-5 fractures was 0.037/1000 exposure hours. As the mean total exposure time for a team of 25 players is about 6000 h/season (around 240 h/player/season), a team can expect an MT-5 fracture every fifth season.

Anthropometric data

The mean age of players sustaining MT-5 fractures was 23±3 (range 18–33) years. This was significantly lower compared with the mean age of the total cohort (23±3 vs 25±5 years, p<0.001). Thirty-two per cent of the players with MT-5 fracture were younger than 21 years and only 22% were older than 25 years. There was no difference in height, weight or body mass index between players with MT-5 fractures and the total cohort.

Primary MT-5 fractures

Among the 67 MT-5 fractures, 45 (67%) were primary fractures and 22 refractures (33%; figure 1).

Injury circumstances

Forty per cent of the fractures occurred during the first 3 months of the season (mainly the preseason period with emphasis on strength and endurance training). Two of three injuries (69%) occurred to the non-dominant (standing) leg. Forty-five per cent of the players sustaining MT-5 fractures had prodromal symptoms (had complained about pain in the lateral side of the foot before the fracture occurred). The majority of injuries (68%) occurred in non-contact situations while the team doctors registered 87% of the injuries as due to trauma and only 13% as due to overuse.

Primary MT-5 fractures with high-quality x-rays available

In seven fractures either images were missing or of low quality, rendering only 38 fractures for further analyses.

Location of MT-5 fractures

Thirty-seven (97%) of the 38 primary MT-5 fractures were located at the base of the bone (zone B in figure 2). Only one avulsion fracture in zone A was seen and no fractures affected the shaft of the bone (zone C).

Figure 2

Anatomical fracture zones of the fifth metatarsal bone (according to Torg et al11). Illustration credit: Vicky Earle.

Treatment and healing rate of primary MT-5 fractures

The avulsion fracture healed and the player went back to full football training after 59 days. Twenty-eight of the 37 zone B base fractures (76%) were treated surgically.

Twenty-one of these healed (healing rate 75%) and returned to full team training at an average of 80±22 (range 43–133) days. Seven players (25%) suffered refractures after a mean of 75±27 (54–113) days. Different surgical techniques have been used but the material is insufficient for an evaluation of surgical methods.

Nine fractures (24%) were treated conservatively. Three of these healed (healing rate 33%) and returned to play full football in 74±13 (range 59–85) days. Five players suffered refractures after an average of 136 (range 32–241) and one player was deemed to have a non-union after 98 days. Surgical treatment had a significantly higher healing rate compared with conservative treatment (75 vs 33%, p<0.05). There was no significant difference in lay-off days for players who were operated on or not (80 vs 74 days, p=0.67; see figure 3).

Figure 3

Flow chart of primary fifth metatarsal fractures (n=38).

Fracture types according to Torg

Seventeen of the 37 (46%) primary MT-5 fractures were classified as Torg type I (clinically acute type, with sharp fracture margins, minimal or no periostal reaction and minimal cortical hypertrophy), and 20/37 (54%) were classified as Torg type II (a fracture line with associated periostal bone union and widened fracture line with a lateral gap, see figure 4).

Figure 4

Roentgenogram of the fifth metatarsal bone, demonstrating a Torg type II fracture with a fracture line with associated-periostal reaction to chronic stress and a widened fracture line with a lateral gap.

None of the fractures were classified as Torg type III (non-union with sclerosis obliterating the medullary canal and blunted fracture edges).

There was no difference in mean age between players sustaining Torg types II and I injuries. Neither was any correlation found between the Torg type and the fracture recovery time, the presence or absence of prodromal symptoms, or whether the injury was caused by stress or trauma or happened in contact or non-contact situations.

As seen in figure 5, the majority of both Torg types I and II were operated on (82% vs 70%). The healing rates after surgery was 86% for Torg type I fractures and 64% for fractures of type II, while the healing rate was low for non-operated fractures (33% for type I and 33% for type II), p=0.052.

Figure 5

Flow chart of primary fifth metatarsal fractures according to Torg types.


This study shows that the MT-5 fracture is a rare injury in football, and with a high rate of healing problems.

Most fractures located towards the base and avulsions are rare

A second main finding is that almost all MT-5 fractures in footballers are located at the base of the bone while avulsions are very rare. This is in contrast with the hospital data about mixed patients, where avulsion of the tuberosity is reported to be the most common location of MT-5 fractures.1 ,2 The review by Kerkhoffs et al6 concluded that avulsion fractures can be treated non-operatively, if undisplaced. However, the transverse fracture through the proximal part or base of the bone, usually referred to as Jones Fracture,12 is reported to be associated with a high rate of healing problems.

MT-5 fractures affect young players

Another main finding, from our study, was that players suffering MT-5 fractures were significantly younger than those who did not. This finding accords with other studies of stress-fractures among athletes.1–3 13–15 Furthermore, we discovered that many fractures occurred at the beginning of the season, especially the preseason training period. Similar findings—that initial symptoms occur during early training or during preseason practice—has been reported by others.1 ,3 ,7

Both these findings support the theory that repetitive stress is a risk factor for MT-5 fractures, since the preseason period concentrates on training which involves repetitive stress. Further, this intensive training period follows a period of rest between seasons: the sudden change-of-loads between these periods might be a risk factor.

There is another factor, that moving from an academy environment—or a lower level of play—to the professional sphere represents intensification. This might explain our finding that the risk of stress fractures is higher among young players.

Prodromal symptoms and x-ray findings suggest bone stress

Another main finding is the presence of prodromal symptoms. In 45% of the cases, pain on the lateral side of the foot was reported prior to sustaining a fracture. Other studies have also reported that prodromal symptoms of pain and discomfort in the area of the proximal fifth metatarsal during sporting activities is common among athletes, suggesting that this fracture could be a stress-fracture.3 ,15 Early recognition can be of vital importance and preventive management at this stage is likely to prevent the injury from progressing. One of the findings from initial x-rays is an area of periosteal bone reaction, a thickening of the cortex and intracortical radiolucency (see figures 4 and 6). This area of radiolucency suggests a pathological process starting at the lateral–plantar cortex at the base of the fifth metatarsal which weakens the bone. An insignificant trauma can subsequently result in a fresh fracture of the medial–dorsal cortex. Further analysis of this pathological process will be presented in a subsequent study. An independent radiologist found bone stress reactions on 54% of the initial x-rays and classified these fractures as chronic or stress fractures. Others have reported similar findings; Josefsson et al2 59%, Torg et al11 46% and Kavanaugh et al3 41%.

Figure 6

An intramedullary screw provides central compression, but does not compress on the lateral–plantar cortex. The lateral gap is not obliterated.

No correlations between Torg type of fracture and injury circumstances

From these findings, it is not clinically obvious whether such fractures should be considered as traumatic or due to stress. Even if the injury occurs in a contact situation, the load in the contact could be below the level that causes a fracture in a normal bone.

Probably the fracture begins as a stress-fracture which weakens part of the bone, with a subsequent and fresh fracture at this level after an insignificant trauma.

Less healing problems are seen after surgery

Treatment options are surgical or conservative, but we found that the union-rate for primary MT-5 fractures among elite-level football players was significantly higher after surgical treatment (75% compared with 33%). Previous studies have reported similar union rates (72–89%) after surgery.4 ,8

Also, we found a union rate of only 33% with conservative treatment, with two players of three showing either refractures or a delayed union. Kavanaugh et al3 reported similar findings with 67% delayed unions among patients with a mean age of 20 years.

However, our finding of a low union rate after non-operative treatment contrasts with other findings, who have reported union rates between 67% and 80% following a conservative treatment.1 ,2 ,8 Studies reporting similar healing rates between surgical and non-surgical treatments of MT-5 fractures are mostly retrospective studies of a mixed patient population where prolonged healing times are frequent.1 ,2 ,8 Our population of young top-level professional athletes are arguably less willing to wait for a prolonged period, and hence more inclined to early surgery. If healing occurs, lay-off times do not differ between the two groups. Return to full sporting-activity averaged 74 (conservative treatment) to 80 days (surgical treatment) and is comparable with the recent literature.15–19

Surgical treatment

Our data were too limited to allow for analyses of different surgical methods. The use of intramedullary screws was the most common method. However, an intramedullary screw provides central compression, but does not compress on the lateral–plantar cortex. The initial radiolucent area is not obliterated (see figure 6). This might explain the relative high percentage of 25% non-union with an operative treatment. Compression on the lateral–plantar cortex can be obtained by means of tension-band wiring (figure 7) or compression-plate fixation, but evidence for the best surgical method is lacking. Addressing the pathological area by means of decortication, bone-grafting or other fracture-healing stimulation might also form an important part of the treatment. ‘Reading’ the fracture is of importance to ensure adequate treatment.

Figure 7

Preoperative and postoperative roentgenograms of a Torg type II fracture. Note the restoration area in the lateral cortex and the subtle cortical reaction (A). Operative treatment with tension wiring and lateral compression (B).

Our study shows a low incidence of these base fractures among professional soccer players. A given club can expect but such a fracture in every five seasons. With a healing rate of 67% (21 of 37 stress fractures healed) this would suggest that such fractures be treated in dedicated centres.

Methodological considerations

The strength of this study is the substantial data about MT-5 fractures obtained from a homogeneous group of male elite-level football players. The majority of publications about sports MT-5 fractures are retrospective analyses from material gathered from hospitals. They usually represent a mix of sexes, of different sports and different performance levels, which make specific-to-sport conclusions difficult. A further strength of the present study is its procedure, with its robust data on lay-off days and possible healing complications.

A limitation of the study was the small number of primary MT-5 fractures with visible fracture lines verified by imaging. This is a problem with all rare injuries. Although data have been prospectively collected from 67 clubs during 11 years, the number of verified primary MT-5 fractures is still limited. Subgroup analyses, which would have been of potentially great interest, were therefore not possible. A comparison of success rates and difference in healing rates between different surgical methods would have been interesting. Comparison of the healing rates of different Torg types, whether treated surgically or non-surgically, would be another example. However the study continues, with larger material, and such subgroup analyses might be possible in the future. Another limitation is that we had no specific instructions for how to define and diagnose non-unions. The diagnosis was up to the team medical doctors. Only one non-union was reported (after 98 days), all the other healing failures were refractures verified by imaging.

In summary, this study shows that attention must be paid to prodromal signs and radiological predictors, because they can alert us, and hence prevent the development of a true MT-5 fracture. If a fracture does occur, surgical treatment is advised with a prognosis of 80 days to full activity.

Owing to the fact that this injury is rare with a high non-union rate, we emphasise the need for dedicated stress-fracture centres. This could benefit both the patient and the scientific evidence in this field of sport injuries, which is currently very scarce.

What are the new findings?

  • Most Metatarsal Five (MT-5) fractures in footballers are located at the base of the bone, while avulsions are rare.

  • The majority of MT-5 fractures are stress-fractures and mainly occur among young players during the preseason high-intensity training period.

  • The stress nature of the injury might explain the high frequency of healing problems.

How might it impact on clinical practice?

  • Less healing problems are seen after surgery compared with conservative treatment.


The authors would like to thank the participating clubs, medical staff and players. The statistical advice of Henrik Magnusson (Msc) is also gratefully acknowledged.



  • Correction notice This article has been corrected since it was published Online First. Please note that Figure 2 has been replaced.

  • Contributors JE designed the study, monitored the data collection, analysed the data and drafted and revised the paper. He is the guarantor. CNvD analysed MRIs and drafted and revised the paper.

  • Funding This study was supported by grants from UEFA, the Swedish Centre for Research in Sports, and Praktikertjänst AB.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval UEFA Football Development Division and the Medical Committee of UEFA.

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

Linked Articles