Do pre-hospital poisoning deaths differ from in-hospital deaths? A retrospective analysis

The study area consisted of the four northernmost counties in Northern Finland. The area had a population of 728,847 inhabitants in 2011. All deaths from poisoning that occurred within the study area between January 1, 2007, and December 31, 2011, were included. Poisoning was defined as fatalities for which the cause of death was registered as ICD-10 diagnosis codes X40-49, X60-69, X85-90, Y10-Y19, T36-T50, T51-T65, Y34, Y57, Y84 [10]. Deceased patients who resided outside Finland were not included in the study as the FCDR only has data on the population residing in Finland. Fatal poisonings were divided into two categories: pre-hospital and in-hospital deaths. Death was defined as pre-hospital if it occurred on scene or during transportation to the hospital, whereas deaths occurring during primary admission to the hospital were defined as in-hospital death.

The death certificates were retrieved for all victims. Medical records were also retrieved for those who died in-hospital. The collected data consisted of patient demographics, the location of death (on scene, during transportation, in-hospital, and other facilities such as jail), intention of intake (suicide, unintentional, treatment failure, or unclear intentions), toxic agents, and involvement of alcohol.

According to Finnish law, medicolegal autopsy is performed if the cause or suspected cause of death is a crime, accident, suicide, poisoning, or unknown [11]. The screenings for alcohol, pharmaceutical products and illicit drugs is conducted in autopsy, if poisoning event is considered possible [12]. In Finland, post-mortem forensic toxicological investigations are centralized at the Hjelt Institute, University of Helsinki.

The statistical analysis was conducted using IBM SPSS Statistics 22 software (IBM SPSS Statistics for Windows, Version 22.0, Armonk, NY, USA). Categorical data are presented as number and percent. Summary measurements are presented as median values with 25th and 75th percentiles and were analyzed using Mann-Whitney’s non-parametric test. Categorical variables were analyzed using Pearson χ².

A total of 2,915 trauma or poisoning deaths occurred in the study area between January 1, 2001, and December 31, 2011, and 689 (23.6%) were fatal poisonings. Most of the deceased were found dead and CPR was attempted significantly less often in victims of pre-hospital deaths than those who died in-hospital (Table 1). The number of annual in-hospital deaths varied from 6 (4.1%) to 11 (8.3%). The victims who died during their hospital stay were more likely to be female and older than those who died pre-hospital.

Most of the victims in both groups were found in the victim’s home (Table 1). Ninety-three of the 647 (14.4%) pre-hospital deaths occurred in a residence of an acquaintance, compared to 3 (7.1%) in-hospital deaths. More information is provided in Table 1.

Ethanol was the most common poisoning agent and was involved overall as the main toxic agent or as a contributory agent in 487 of the 647 (75.3%) pre-hospital deaths and 19 of the 42 (45.2%) in-hospital deaths (p < 0.001). Ethanol was the causative agent significantly more often in pre-hospital deaths than in-hospital deaths. For the in-hospital group, the second most common agent was methanol (Table 2).

Of opioids, tramadol was found as the primary toxic agent in 31 (4.5%) of the post-mortem forensic toxicological screenings, codeine in 27 (3.9%) and buprenorphine in 12 (1.7%). There were no illicit street drugs found among the primary toxic agents of fatal poisonings. As a secondary toxic agent, amphetamine was found in 11 (1.6%) fatalities and cannabinoid agent in 10 (1.5%), but apart from those findings there were no signs of illicit street drugs. Heroin wasn’t found in any of the fatalities.

In suicides, ethanol was the main agent in of the 15 of the 122 cases (12.3%) followed by amitriptyline (11 (9.0%)), codeine (10 (8.2%)) and insulin (8 (6.6%)).

Multiple ingestions occurred more often in pre-hospital deaths than in in-hospital deaths (52.2% vs. 35.7%, p < 0.001). Most common combination included in multiple ingestions was ethanol and diazepam (N = 50), and generally the most common combination was ethanol combined with benzodiazepine (N = 136) or antidepressant (N = 83). The most common combination in co-ingestions where ethanol wasn’t involved was tramadol and diazepam (N = 16).

The main limitation of the present study is the low number of in hospitals-deaths, which partly limited the possibilities for comparing poisoning patterns between pre-hospital and in-hospital deaths. In addition, the data available from hospital stays were limited because of the retrospective design of the study. Furthermore, the suicide rate may be underestimated in the present study, as the intention of ingestion cannot be determined reliably, even in medicolegal autopsy. Suicidal intent is recorded only if there is clinical evidence supporting this, such as good-bye letter. However, with an autopsy rate of 100%, we present the most reliable results that can be reached in this population. Due to the high quality of death certificates after medicolegal autopsy, the location of death was definitive, as well as the pre-hospital emergency critical care.

Considering that the coverage of the cause of death registry is nearly 100% [15], the gathered data should accurately reflect the quality of poisoning deaths in the source population.

According to the present results and previous literature, prevention seems to be the most important factor when aiming to reduce the number of fatal poisonings in Finland; a majority of the victims are not reached early enough and non-survivors who are hospitalized are poisoned by methanol or are severely ill. Though the proportion of deaths resulting from ethanol poisoning was high, access to psycho-pharmaceutical agents is important as well as the prevention of alcohol use. The most common place where victims were found or the poisoning occurred was a residence, and only a few of the poisonings occurred in locations where the event could have been prevented, including imprisonment or health care facilities. Most of the victims were found dead and the possibilities for decrease the rate of poisoning deaths with advances in pre-hospital emergency care or critical care are limited.

In light of the present results, the most effective way to decrease overall poisoning-associated mortality in the source population seems to be preventing alcohol intoxication. Recognizing existing risk populations for alcohol abuse is one step in preventing alcohol intoxication and fatalities [16, 17]. In Finland, fatal alcohol poisonings seem to peak around weekends and festive events; a 1% increase in alcohol sales increases the number of fatal alcohol intoxications by 0.4% [18]. On a societal level, policy changes, such as banning alcohol advertising, making alcohol more expensive, and reducing availability, have been proven to be effective in reducing harm caused by alcohol, whereas increasing education and information on the harm caused by alcohol has not been shown to have as great of an effect [19]. Furthermore, removing methanol-based antifreeze solutions from the market should be considered. Upon joining the European Union in 1995, previously prohibited antifreezes containing methanol became available for sale in Finland. This resulted in a remarkable increase in fatal methanol poisonings in subsequent years [20].