Drowning deaths in Sweden with emphasis on the presence of alcohol and drugs – a retrospective study, 1992–2009

Annual incidence of drowning deaths for different age groups, males, and females were calculated and presented as cases per 100,000 of mean population. A Poisson regression model was used to analyse time trend, age effects, and differences between males and females. Results from the Poisson regression model were given as incidence rate ratios (IRR) and 95% confidence intervals (CI). Comparison of differences between the groups unintentional and intentional drowning with respect to alcohol use and sex was done using the Chi-square test performed in Epi Info 3.4. SPSS 19 was used for other statistical calculations.

Unintentional drowning caused 2,585 (50%) deaths – 2,180 (84%) were males and 405 (16%) females. The mean age for males was 54 years (range 1–101, SD 22), and the mean age for females was 56 years (range 0–93, SD 26). The highest numbers of drowning deaths were in age groups 50–59 and 60–69 years (17% and 18%, respectively). Most unintentional drowning occurred in lakes, sea, streams or rivers, or bathtubs (Table 2). Thin ice was associated with 262 (13%) drowning incidents that occurred in lakes, seas, streams, and rivers.

Unintentional drowning deaths were associated with bathing and other unspecified water activities (1807, 70%), boating incidents (521, 20%), cars driving into water (82, 3%), snowmobiles (78, 3%), diving (51, 2%), and airplanes, helicopters, and other vehicles (46, 2%).

In 909/5,125 cases (18%), intent was unable to establish. This group mainly consisted of males (623, 68%). The mean age for males was 53 years (range 0–93, SD 19) and the mean age for females was 59 years (range 0–93, SD 20). Incidence increased with age for both males and females (Figure 1). Stream, river, bathtubs, sea, and lakes were the most frequent drowning locations. Significantly more females drowned in a bathtub compared to males also in this group (p<0.01) (Table 2).

In total, 4,377 out of 4,812 (91%) individuals were tested for alcohol and 1,656 (38%) had BAC ≥ 0.2 g/l in blood with a mean BAC of 1.8 g/l (range 0.2-5.2). The presence of alcohol was more frequent in unintentional drowning deaths compared to suicidal drowning deaths (p< 0.001) (Table 2) and in males compared to females (p<0.01) (Table 3).

In all, 4,181 out of 4,812 (87%) individuals were tested for pharmaceutical substances in their blood (Table 3). In 1688 (40%) individuals, one or several psychoactive substances were detected. Multiple psychoactive drugs were found in 1,070 (26%) individuals. The most common psychoactive drugs were benzodiazepines and antidepressants (Table 3). The individuals who committed suicide had the highest proportion of pharmaceuticals compared with the undetermined and the unintentional groups (p< 0.001) (Table 5). The individuals between 50–79 years of age had the highest proportion of psychoactive substances (57%) in their blood (p<0.001). In 82 (10%) of 854 tested individuals, illicit drugs were detected in the blood (Table 3). A higher proportion (15%) of illicit drugs were found in the undetermined group compared to unintentional drowning deaths (9%) (p<0.001) (Table 5). The most common illicit drugs were amphetamine 42/82 (51%) and tetrahydrocannabinol 34/82 (41%).

A substantial proportion (44%) of unintentional drowning deaths tested positive for alcohol in the blood. This finding could be compared with 51% in Finland [5], 35%-55% in the USA [40], 50% in New Zealand [41], 62% in Ireland [32], and 22% in Australia [13]. The proportion of alcohol in drowning deaths may reflect the differences in alcohol policy and consumption in these countries.

In our study, 54% of boating fatalities were alcohol positive, a percentage that is comparable with a study from the USA [42]. The relative risk of dying in boating incidents increases with higher blood alcohol level and is increased 16-fold at a BAC of 1.0 g/l [42]. Legislation may affect the use of alcohol when driving a boat. In Sweden, until 2010 the legal limit for alcohol in drivers of larger boats (greater than ten metres) was 0.5 g/l; later this limit was lowered to 0.2 g/l, which is the same as for drivers of motor vehicles in road traffic. At present, there is no limit for alcohol in drivers of smaller recreational boats (less than ten metres) although it seems obvious this should be implemented.

For road traffic, it is well known that alcohol impairs a person’s judgment, performance, and behaviour [43, 44] and increasing levels of impairment are associated with increased BAC levels [44]. Moreover, alcohol affects coordination and increases the risk of hypothermia [45]. Therefore, reducing alcohol use in combination with water activities may prevent many of these deaths [46].

In the present study, alcohol was found in 24% of suicidal drowning and alcohol may also have facilitated suicide. It has been reported that alcohol lowers the threshold for suicide; the presence of alcohol was 23% for suicidal drowning death compared to 38-64% for more violent suicide methods [47]. Consequently, heavy drinkers with a history of mental illness might be at higher risk of suicide [35, 48].

Almost 25% of those who drowned had one or more pharmaceuticals in their blood. Benzodiazepines and antidepressants were most common, especially in the older age groups. Psychoactive drugs affect cognitive function, concentration, vision, coordination, and balance depending on brain concentration, individual susceptibility, and interaction with other drugs [49]. These drugs may affect risk-taking behaviour as well as affect survival possibilities after falling in the water, i.e., swimming ability. For boating, psychoactive drugs may play a similar role as for drivers in road traffic where benzodiazepine and antidepressants have been correlated with increased crash risk [50, 51]. Those who committed suicide had the highest percentage of psychoactive drugs, which probably is related to a higher psychiatric morbidity. The combination of alcohol and psychoactive drugs might have an additional effect on cognitive functions [52]. However, it is difficult to know how individuals are affected as this depends on interaction with other drugs and whether higher dosages than therapeutic recommendations were used. In this study, illicit drugs were detected in almost 10% of the cases and it was higher in the undetermined group. However, most of the drowning deaths were not tested for illicit drugs. More extensive testing is needed to obtain more reliable data to assess the role of illicit drugs in drowning deaths.

The findings in this study suggest that it is important to reach middle/older age groups when designing future prevention programs, a recommendation also suggested by American and Australian researchers [13, 53]. Education, knowledge, and training are important factors. The information about the danger of combining alcohol and drugs with water activities should be given to groups at high risk, i.e., males and the middle/older age groups. Eventually, it has previously been found that it is more difficult to protect adults than children from drowning [7]. Prevention strategies for adults should include a focus on swimming ability, ice prods, and personal floatation devices that could affect survival (i.e., the time in water). Immediate rescue and cardiopulmonary resuscitation by bystanders are also important factors that affect survival [54]. It is also important to emphasise passive measures such as designing bridges in such a way that persons cannot reach the edge, making it more difficult to jump or fall from a bridge.

To our knowledge, this is the first comprehensive study of drowning in Sweden. As the autopsy rate is high in Sweden for drowning deaths, probably only a few cases were missed. All autopsied cases of drowning in Sweden were analysed, including unintentional, intentional, and undetermined drowning. This approach provides a complete data set. As this type of thoroughness is often not the case in other studies, it is difficult to compare the figures with previous studies. In our study, we have information about alcohol, pharmaceutical drugs, and illicit drugs in femoral blood, whereas many previous studies on drowning death do not provide data on drugs [55].

Of all the drowning deaths, 18% of the cases were undetermined. According to a previous study that analysed the coding practices of injury deaths in Nordic countries, Sweden had a higher rate of undetermined cases and many of these were water related [24]. Therefore, the coding practice might underestimate drowning suicides in Sweden.

The data on alcohol in drowning deaths should be used with caution. Post-mortem microbial activity can produce misleading alcohol levels [18]. To avoid overestimation of alcohol, we have excluded decomposed bodies from alcohol analyses and the cut-off level for alcohol was set to 0.2 g/l in all cases. Therefore, our data represent a conservative estimate. Another way to limit this problem is to perform blood sampling within 24 hours after immersion [56], but this was not possible in all cases. It is difficult to compare studies since information about the proportion of decomposed bodies and source when obtaining the blood sample for alcohol measurement and toxicology screening often are missing. In the present study, most of the drowning deaths were not tested for illicit drugs, so there might be a selection bias.

Because the present study was based on register data, information regarding some circumstances could not be obtained, i.e., pre-drowning activities, swimming skills, use of personal floatation devices, and rescue attempts.