METHODS

English language literature published up to October 2003 was searched for relevant articles. The main keywords used were “drown”, “drowning”, “immersion”, “alcohol”, “drink”, and “drinking”. Initial searches were through PubMed⁷ and the library of the Research Centre for Injury Studies. Secondary follow up of sources cited in reference lists of obtained papers also provided a number of appropriate papers.

Recreational aquatic activities were defined as activities explicitly related to water that are undertaken for fun, pleasure or amateur sport, and included swimming, surfing, boating, water skiing, underwater diving, and fishing. Any relevant setting was included, such as oceans, beaches, rivers, creeks, lakes, and public and private swimming pools. Drowning deaths that occurred as result of swimming activity were examined separately to those during boating. Deaths that occurred when the contact with the water was incidental (such as falling in while walking near water), or the activity was not recreational (such as washing in a bathtub or working on a vessel), were excluded as much as possible. In practice, this was usually possible if enough information was included in the paper, but it was sometimes difficult to exclude all unwanted cases because of the way the data were presented in the available papers and reports. Similarly, it was sometimes difficult to confidently identify a subgroup of cases that only included swimming. Relevant methodological aspects of the study, including the extent to which the relevant subjects were only engaged in recreation, the definition used by the study of the activities equivalent to “swimming” or “boating”, the age of the subjects included in the part of the study relevant to consideration of the involvement of alcohol, the number of subjects in the relevant group, and the number (and proportion) of subjects with valid blood alcohol readings, are included in the study summaries presented in tables 1 and 2.

METHODOLOGICAL CONSIDERATIONS

Physiological and psychological effects of alcohol

Physiological and psychological effects of alcohol have potential to increase the risk of immersion injury for persons involved in swimming and other water related activity because they increase the likelihood of exposure to high risk situations, and the harm that might result, by affecting cognitive processing, central nervous system processing, and physiological responses. In particular, impaired judgment increases the likelihood of persons choosing to be, or not avoiding being, exposed to higher risk situations; blood vessel dilatation increases the period of time someone may choose to remain in cold water, thereby increasing the risk of sustaining hypothermia; and labyrinthine dysfunction and laryngospasm decrease the chances of survival if in the water. These factors can also increase the difficulties facing any potential rescuer.^5,8–12

Assessing the contribution of alcohol

Several factors need to be considered when assessing the evidence for a causal or contributory role for alcohol in drowning associated with recreational aquatic activity. A key consideration is whether alcohol relatedness/involvement refers solely to the “presence of alcohol”, or to whether alcohol “contributed to” or “caused” the injury event because of impairment of relevant parties. Confounding factors (for example, polydrug use, how learned the activity is, hangover effects, and psychophysiological consequences) complicate the relationship between intoxication and impairment further. “Alcohol relatedness” can also refer to alcohol use by other parties involved in the injury event, such as a boat operator when a passenger drowns, although information on this type of alcohol relatedness is rarely collected routinely.

It is, in principle, possible to assess injured people, case by case, to determine whether alcohol might have contributed to the occurrence of each injury case. That can be done by obtaining evidence that the injured person, or some other person involved in the event that resulted in injury, had a significant level of blood alcohol at the time of occurrence, and that the circumstances of the incident were such that it was likely that the level of intoxication of the person contributed to the injury occurring and/or to the resulting injury being more serious than it might otherwise have been. However, the necessary level of information is often lacking.¹³

Another important consideration is the fact that blood alcohol levels can rise after death due to decomposition processes, or fall before death due to normal metabolic processes. Therefore, it is important to obtain information regarding when the blood used for blood alcohol level estimation is taken in relation to both the incident and the death.^14–17 Various criteria for this have been suggested.^18,19

RESULTS

Use of alcohol in recreational aquatic activities

Most information on the use of alcohol in association with recreational water activity comes from telephone surveys or face-to-face interviews conducted in the United States, either at a national or state level. These surveys have consistently suggested that about 30%–40% (higher in some studies) of people on boats drink alcohol while on board,^10,20–24 that men tend to drink more^{10,20–22,24,25} and behave in higher risk aquatic activities in association with drinking,²² than women, and that available boater training is inadequate.^{10,11,20,25–28}

There is little published information about public attitudes or understanding of alcohol’s involvement in serious and fatal injuries sustained during recreational aquatic activity, with one United States study finding that many respondents to a mail survey reported that it was safe for passengers to drink more than boat operators, and that it was safe to drink more when the boat was at rest compared with when it was moving.²⁶

Contribution of alcohol to recreational aquatic activity injury

In their review of alcohol and drowning literature published up to 1985, Howland and Hingson found that between 25% and 50% of adult persons who drowned unintentionally had used alcohol, but they also called for more information to be collected on alcohol use in involved persons who did not drown.⁵ A subsequent review by the same authors of literature up to 1991 found a similar range of values.⁴ Several relevant studies have been published since then, and their results remain consistent with the Howland and Hingson reviews.^2,29–31

Most potentially relevant studies are not restricted to persons who were injured in the context of recreational aquatic activities. A recent review does focus on recreational aquatic activity, but not on drowning or alcohol, although these are included.⁶ The quality of information on alcohol involvement in individual studies ranges from poor to very good. These studies suggest that drowning is the overwhelming cause of death associated with recreational aquatic activity, that alcohol is detected in the blood of about 30%–70% persons who drown in association with recreational aquatic activity, and that alcohol probably contributes to between 10% and 30% of all recreational drowning deaths. The useful results from the more relevant of these studies are summarised in table 1 for swimming (including surfing),^2,3,32–39 and table 2 for boating.^{28,32–36,38–44}

Measures of risk due to alcohol

Very few direct measures of risk of serious or fatal injury related to aquatic recreational activities are available. Those that are available nearly all cover only drowning. The main available estimates are described here and summarised in table 3.^{28,42,44–46}

The best available information comes from a population based case referent study of drinking and recreational boating fatalities by Smith and co-workers. This study appears to have included all fatalities, but these were predominantly drowning deaths. The study found that alcohol use by the deceased person increased the risk of fatal injury associated with boating at all measurable levels of alcohol, ranging from an estimated relative risk (RR) of 1.3 (95% confidence interval (CI) 1.2 to 1.4) at 0.010 g/100 ml, 10.6 (95% CI 6.9 to 15.7) at 0.10 g/100 ml to 52 (95% CI 26 to 106) at 0.25 g/100 ml. The relative risks were similar if fatalities (and the relevant controls) of people who were voluntarily swimming from a boat were excluded, and were similar for boat operators and boat passengers.²⁸

The only other comparable study was a small study of boating fatalities in California, which found an odds ratio of 2.9 for any drinking, and 10.6 for blood alcohol levels above 0.10 g/100 ml (Mengert et al⁴⁵ cited elsewhere^26,28). The remaining studies were either small⁴⁶ or provided only absolute measures of risk^42,44 and appeared to have considerable methodological limitations.

Trends in the role of alcohol

Cummings and Quan applied the attributable percentages they derived to cases identified in coroners’ records with non-zero blood alcohol levels in Washington State, in the United States. The study found that there was an 81% decline in drowning due to alcohol over the 20 years covered by the study, and that the percentage of all drowning due to alcohol fell from 50% to 22% over the 20 years.³¹ The reasons for the decline have been debated, but might include any of a decreased propensity to swim in settings where alcohol is more likely to be used, decreased use of alcohol, and bias due to missing data on alcohol.⁴⁷ A similar study of the trends in the role of alcohol consumption and fatal “accidents” (including drowning) in Canada from 1950 to 1998 found “statistically significant” positive relationships between per capita alcohol intake and fatality rate for “falling accidents”, “motor vehicle accidents” and “other accidents”, but not for “drowning accidents”.⁴⁸

Prevention strategies

Many strategies have been suggested for decreasing the risk of alcohol related injury (particularly drowning) during aquatic activities. These are contained in many of the references already cited and are shown in table 4.^{2,3,5,10–12,20–28,33–36,39–41,43,44,49–61}

However, few of these suggested strategies have been the subject of rigorous scientific evaluation, and drowning has been called the “final frontier” of injury prevention.⁶² Some intervention trials of limited relevance to alcohol and drowning have been published. All parents of children attending an emergency department at a paediatric tertiary referral hospital in the United States who were given printed information (which included mention of alcohol) on water safety recalled receiving instructions of some sort when surveyed one to two weeks later.⁶³ Minimum drinking age laws do not appear to be an important policy initiative for preventing drowning in adolescents and young adults, based on a study of the number of drowning fatalities and the state minimum drinking age laws in the United States, although the authors did not rule out a small effect of the laws.⁶⁴

Chochinov, in a review of boating fatalities in Canada, argues that alcohol is an important contributor to boating related drowning, and that the use of alcohol in association with the operation of a boat should be seen as socially irresponsible, as is drinking and driving.⁵² However, the focus on boat operators (“designated operators”) in campaigns to decrease alcohol use in relation to boating activity has been criticised, because drinking passengers may be at risk regardless of the sobriety of the boat operator through the effects of alcohol on balance, judgment, swimming ability, and response to cold.^11,26,28 The authors also argue against focusing only on moving boats, since many boating fatalities occur when the boat is stationary or drifting: 24% in the United States in 2000.⁴²

In addition to the suggested prevention areas listed above, a number of areas for further research or action have been identified (table 5). These initiatives can be combined into a small number of themes, such as initiatives to discourage use of alcohol; impose penalties on persons who have high blood alcohol levels; focus on particular high risk groups; and obtain a better understanding of the risks of alcohol use and the benefits of prevention programs.^{3,10,20,21,29,31–33,37,65}

DISCUSSION

Activities designed to decrease alcohol consumption in relation to recreational aquatic activity are based on a presumption that alcohol is a risk factor, and are thus generally accepted as being appropriate preventive activities. Anecdotal evidence and cases series studies strongly support the presumption that alcohol use is a problem, but it is really only the case referent study by Smith and co-workers of fatalities related to recreational boating that provides robust evidence of the extent of increased risk associated with various levels of blood alcohol.²⁸

The overall quality of the studies, and in particular their usefulness in the current consideration of drowning in recreational aquatic activities, varied considerably. The most useful studies had a study group that was a close match to the groups of main interest (recreational swimming or boating), clearly stated the subject numbers and the number with valid blood alcohol measurements, had a high proportion of subjects with valid blood alcohol measurements, clearly stated the criteria for inclusion and exclusion of blood alcohol measurements, and provided numeric values for the blood alcohol measurements. There was insufficient information to provide useful insight into the involvement of alcohol in areas of aquatic recreation apart from swimming and boating resulting in drowning, such as underwater diving and rock fishing, or into any possible differences with age and gender, apart from the apparent tendency of men to drink more alcohol in relation to aquatic activity, as mentioned earlier.

The rate of autopsy in drowning varies widely between regions, as does the proportion of persons for whom blood alcohol levels are assessed, either electively or as mandated by legislation. Most of the studies reviewed here had a significant minority, or even a majority, of cases with missing blood alcohol values. Where the percentage was known, valid blood alcohol values were usually available for between 50% and 80% of subjects. A lack of valid blood alcohol values, either because the body was not recovered, the body was significantly decomposed, the body was not recovered for several days or the specimen was not taken at autopsy, can potentially bias any study in this area. This is because particular drowning circumstances may be more likely to both involve alcohol and to result in long delays in recovery of the body.¹⁹ Therefore, it is important to minimise the number of subjects with missing blood alcohol results. Since a considerable proportion of subjects in many studies would need to be excluded if the blood specimen needed to be taken within 24 hours, as recommended by some,¹⁸ techniques for adjusting blood alcohol measurements in bodies found between 12 hours and seven days after death have been used.¹⁹

PRIORITIES FOR FUTURE WORK

Information about the use of alcohol in the context of aquatic activities is inadequate for most countries for reliable estimation of the proportion of drowning attributable to alcohol, or to provide a basis for information based policy response (for example, monitoring effects of programs to discourage drinking before swimming or while boating). Effective prevention is likely to result from a thorough understanding of aquatic activities, alcohol use, and the way that these interact. Several forms of information can contribute to this understanding. Case-control studies can enable comparison of characteristics of participants in aquatic activities (including their use of alcohol) who did and did not drown (or experience another undesirable outcome). Surveys of participants can improve understanding of patterns of use of alcohol during aquatic activities, and about the knowledge and attitudes of participants to such use. Trials can be designed to test and evaluate the effectiveness of interventions designed to reduce alcohol related harm during aquatic activities. In addition, better surveillance at the population level is required, especially in developing countries, of the extent and level of alcohol involvement. This surveillance requires a high proportion of drowning cases to have autopsies, and a high proportion of drowning cases to have valid blood alcohol measurements.

The existing evidence, argument by analogy with literature on risk due to alcohol in other settings (for example, motor vehicle operation), and a cautious approach provide a basis for arguing against any use of alcohol during, or in relevant periods before, aquatic activities. However, the specific evidence base warrants expansion, particularly to confirm apparent elevation of risk at low blood alcohol levels (that is, 0.05 g/100 ml and lower) and to broaden the range of types of aquatic activity for which evidence on alcohol and risk is available. Existing studies provide relevant examples of how such investigations could be undertaken.^28,46 Table 5 lists a number of possible prevention initiatives. These and other approaches, such as legislation using the same blood alcohol limits for boat operators as for motor vehicle drivers, training in responsible serving of alcohol (on and off-shore), availability of lifesaving equipment, changes in practice for night-time aquatic activity, and policy/practice changes for operators of marine facilities and licensed venues, all appeal as potentially effective. However, without proper evaluation studies of interventions, it is difficult to confidently recommend one or more as likely to be effective in decreasing the contribution that alcohol makes to drowning.

CONCLUSIONS

Alcohol appears to be widely used in association with recreational aquatic activity, and to be an important risk factor for drowning associated with recreational aquatic activity. Information is lacking in many areas concerning alcohol use in relation to recreational aquatic activity. The most important of these relate to understanding the attitudes and alcohol use of persons engaging in recreational aquatic activity, characterising the increased risk (and the attributable risk due to alcohol) of drowning that arises from different levels and types of alcohol use, and evaluating the many prevention approaches that have been suggested or already implemented.