Intended for healthcare professionals


Preventing injuries from bar glasses

BMJ 1994; 308 doi: (Published 09 April 1994) Cite this as: BMJ 1994;308:932
  1. J Shepherd

    The Home Office has estimated that each year in Britain between 3400 and 5400 offences occur in which glass is used as a weapon. Two surveys of victims of assault who attended accident and emergency departments in Bristol and south London found that the most commonly used sharp weapons were bar glasses.1,2 Another study found that three quarters of such injuries had arisen through assaults with straight sided bar glasses of one pint (0.57 1) capacity (“noniks” - or no nicks).3

    Contrary to expectation, three quarters of the glasses were intact until they were thrown or thrust at someone and then broke on impact. Almost all injuries were to the face, and doctors working in accident and emergency departments predicted that deformity at six months would be “noticeable” or “very noticeable” in three quarters of the victims. The British Association of Hand Surgeons identified more than 200 accidental hand injuries due to bar glasses in three months in 1987.4 These surveys highlight the morbidity produced by bar glass - mostly in young people, in whom initiatives aimed at preventing accidents have been given priority in the Health of the Nation.

    Glasses' resistance to impact varies according to manufacturer and degree of wear and tear. A laboratory investigation of new and worn glasses of one pint capacity (nonik and tankard designs) available in Britain found that noniks from one manufacturer were more than six times more resistant to impact than all the others, and when they were worn they were twice as resistant to impact as similarly worn noniks from other manufacturers. These noniks were also three times more resistant to impact than tankards. Heavy wear and tear, however, substantially weakened all the designs.5

    When the glasses that were comparatively resistant to impact failed they disintegrated into cuboid fragments with angles that tended towards 90°. This particularly applied to the thicker bases. In contrast, other glasses disintegrated into larger, jagged pieces, and the base of the glass generally remained intact and usable as a weapon.

    In the search for an explanation for these differences it became apparent that the glassware that was resistant to impact had been tempered (toughened) during manufacture while all the other glassware had been annealed. Although tempering has long been applied to the manufacture of car windscreens, plate glass, and cooking containers, only two manufacturers currently temper bar glasses. This process involves rapid cooling of the glass after its initial formation so that a compressive outer skin is produced. This holds together the outer layer of glass and particularly the microflaws and cracks that are common to all glassware after manufacture. This explains both their increased resistance to impact and the different pattern of fragmentation.

    As well as its increased safety, tempered glassware has several other advantages, including durability and longevity. For example, in a large office complex tempered glasses lasted up to 25 times longer than annealed glasses of the same design.6 Manufacturers that do not produce tempered bar glassware claim that injury may follow explosive disintegration of tempered glassware and that such glassware discolours more than annealed glassware,7 though no evidence for this has been published.

    Following these findings, a search elicited no safety guidelines or codes of practice in relation to bar glasses in Britain or internationally. Nothing therefore prevents manufacturers from producing glasses of thinner and thinner material; indeed, a commercial incentive exists to do so, particularly as the British market for bar glasses is worth about pounds sterling100m a year.

    A survey of bar workers found that 40% had been injured by bar glasses - mostly while stacking and washing noniks of one pint capacity.8 Although most injuries were to the hand and produced only minor inconvenience, about one in 10 substantially disrupted work and one in three required treatment in an accident and emergency department. Clearly, this level of injury from sharp objects would be unacceptable in a laboratory or other working environment.

    What should be done? Firstly, safety standards need to be developed to protect consumers and bar workers. Glassware in many countries bears a government stamp to signify that the capacity of the glass conforms to certain standards of volume. This should also signify adherence to standards of safety. In Britain an obvious objective is that every half pint or one pint beer glass should bear the British Standards Institution's “kite mark.” Secondly, a code of practice should be developed regarding the type of glass that should be used in particular environments - for example, in pub gardens and urban pubs. Thirdly, the safety of disposable and reusable plastic containers should be acknowledged. Although levels of injury do not warrant wholesale replacement of glass, plastic has obvious advantages in some environments - for example, at beer festivals and sporting occasions. And lastly, licensees should ensure that empty bottles and glasses are regularly collected and that worn glasses are replaced before they break in use.9

    Jonathan Shepherd's group received a research grant to study risk factors for urban violence from the glass manufacturers, J G Durand. This followed the publication of the results of a study by Shepherd and colleagues reporting that tempered glasses - as made by Durand - were more resistant to impact than annealed glasses (BMJ 1991;303:1330).


    View Abstract