Spinal cord injuries in rugby union players

How much longer must we wait for proper epidemiological studies?

Recognition nearly 20 years ago of a high incidence of spinal cord injuries in rugby union players1 led eventually to changes in the rules in most rugby playing countries. New Zealand changed its rules in 1980 and 1984; Britain in 1979, 1984, and 1985; Australia in 1985; and South Africa in 1990. The consensus is that these changes have produced the desired effect, but how robust is the evidence? The week in which the Rugby World Cup begins in South Africa seems an apt time to address this question.

Silver's study of 63 spinal cord injuries in rugby union players between 1952 and 1982 was important in identifying mechanisms of injury in a representative sample of rugby players.2 But it was not a study of incidence. Two further studies by him included all spinal cord injuries in rugby players treated in eight spinal units in England and Wales between 1980 and 1986.3 4 By this time the incidence of spinal cord injuries should have begun to fall if the actions taken by the English Rugby Football Union had been effective.

These data show that, although the annual incidence of spinal cord injuries at Stoke Mandeville Hospital fell steadily from a peak of nine admissions in the 1980-1 season to two in 1986-7, the overall incidence of reported spinal cord injuries in rugby union players in England and Wales remained fairly constant between 1980 and 1986. Clearly these data do not support Silver's conclusion that a fall in the incidence of serious injury in Britain has followed changes in the laws of rugby.5 Until complete data from all spinal units in Britain have been analysed (the numerator) and the total number of rugby players in Britain is known (the denominator), no conclusions regarding the effects of recent rule changes can be drawn.

Similar uncertainty exists in other rugby playing countries. Burry and Calcinai reported that the average annual incidence of spinal cord injuries in New Zealand fell from nine between 1973 and 1978 to less than three in 1985-6 as a result of specific rule changes that reduced the number of injuries in scrums and loose rucks or mauls.6 Calcinai's more recent data show that the number of spinal cord injuries in New Zealand rugby players rose again in the 1988 and 1989 rugby seasons: incidence in 1989 was not lower than that before the rules were first changed in 1980.7 (Only three further injuries were reported in the 1990 and 1991 seasons so the average annual incidence between 1988 and 1991 fell back to four.)

But the accuracy of these data must now be questioned in the light of Dixon's analysis of hospital discharges for rugby injuries, including spinal cord injuries, in New Zealand rugby players between 1980 and 1990.8 If we assume that Dixon's methods were correct Calcinai and Burry apparently identified only about half of all spinal cord injuries in New Zealand rugby players.

Interestingly, Dixon showed that spinal cord injuries peaked in 1984 and 1986—after the introduction of what were considered to be effective rule changes. It is therefore premature to conclude that the incidence of spinal cord injuries related to rugby has fallen in New Zealand.

The sole data for spinal cord injuries in South African rugby players come from a 30 year survey of admissions to the spinal unit at the Conradie Hospital in Cape Town, the source of the original study of spinal cord injuries in rugby players.1 9 The most recent analysis of those data shows that admissions of rugby union players with spinal cord injuries rose progressively during the 1980s with a peak incidence of 12 admissions in 1989.10

In the past four seasons (1990–93), there have been a further 31 spinal cord injuries in rugby players. This is fewer than the 37 such injuries in the four preceding seasons, owing mainly to a 33% reduction in the incidence of spinal cord injuries in schoolboys after rules were changed for under 19 rugby at the start of the 1990 season. Hence the evidence from South Africa is that the incidence of spinal cord injuries has altered recently only in schoolboy rugby players.

Assessment of the effects of rule changes in Australian rugby depends on follow up data to Taylor and Coolican's study,11 which have yet to be published. An official publication of the Australian Rugby Football Union, however, notes that since the introduction of rule changes to schoolboy rugby, no serious spinal cord injuries have been reported in Australian schoolboy rugby players in the past eight seasons (1985–92).12

Awareness is now growing of the lack of data on “near miss” injuries, in which serious cervical injury occurs without damage to the spinal cord. Calcinai found that whereas his method of analysis identified only one spinal cord injury in New Zealand rugby players in 1991 there were at least 10 “near miss” injuries.7 Similarly, in a prospective one year study of rugby injuries at 25 high schools in the Cape Province of South Africa, cervical ligament injuries were reported, seven of which were associated with cervical fracture or dislocation but none of which caused cervical injury.13 Analysing only those injuries that cause spinal cord injury therefore identifies only a small fraction of the total number of cervical injuries in rugby players. The mechanisms causing near miss injuries are similar to those causing spinal cord damage7 13; hence effective interventions to prevent paralysing injuries would also reduce the incidence of these injuries.

Nearly 20 years after the BMJ first drew attention to the issue we still do not know the true incidence of either spinal cord or all cervical injuries in rugby players in any rugby playing country. This unacceptable situation will continue until accurate data are collected, an idea that editorials in this and other journals have been promoting for years.1 14 15 The international community of doctors involved in rugby must convince rugby administrators in all countries to set up epidemiologically valid surveys of injuries. Until they do so there will not be enough accurate data to support change.