Using a large cohort of over 100,000 ICU patients, this study demonstrates that intoxicated individuals have a better short-term outcome than patients with other critical illnesses regardless of initial lactate, as well as various baseline (age, sex, ethnicity), and patient care characteristics (level of care, mechanical ventilation, vasopressor use). Since intoxications can be caused by a vast variety of substances and poisoned patients present with a wide range of clinical manifestations, they have always been a challenge for acute care physicians. Despite the etiologic diversity and relatively high general prevalence of intoxicated patients in ICUs, the need for invasive measures is comparatively rare. This analysis is intended to provide further important data regarding specific patient characteristics of intoxicated patients, and in particular those individuals who have a poor outcome, in order to help to identify them early in terms of better resource allocation.
In general, we observed a significantly lower mortality of intoxicated patients compared to the entire remaining collective. It is certainly not without reason that “intoxication” is described in current resuscitation guidelines as "reversible cause" and usually represents an acute event with comprehensible pathophysiology and clinical course, frequently also without relevant pre-existing (chronic) organ damage [12]. With regard to outcome parameters, we found "less ill" patients in the intoxication group. Despite from similar rates of dialysis in the first three days, less patients suffered from AKI in the group of intoxicated patients. This is in accordance to our clinical experience, since the elimination of many substances can be accelerated with hemodialysis and/or –filtration [13]. A somewhat surprising finding is the higher proportion of women with acute poisoning, as a greater proportion of men in this patient population is frequently described in the literature [14]. A possible explanation could be intentional overdose (which is observed more frequently in women in contrast to e.g. exposure to chemicals) and high rates of sedatives as causative substances in our population which are used more often by female individuals [14]. However, Brandenburg et al. observed a similar finding in their large-scale analysis on intoxicated patients, as well as a likewise relatively young average age of patients [2]. Also, male patients had a numerically worse outcome compared to their female counterparts. This finding cannot be explained causally on the basis of the available data. Intoxicated patients in our study had lower BMIs, whereas underweighted patients within this group had worse outcomes. In the past a higher mortality in underweighted individuals has been observed in several medical conditions, as well as in the general population [15, 16]. In contrast, obese patients often show better outcomes, which is also known as the “obesity paradox”. The causes for weight-related outcome differences are complex and range from severe pre-existing chronic diseases, different drug distribution patterns of especially lipophilic substances, over nutritional status, socioeconomic factors up to immunological phenomenons [15,16,17]. For an accurate assessment of BMI-related outcomes, a standardized nutritional assessment, a detailed analysis of pre-existing conditions, but also a functional analysis regarding activities of daily life (ADLs) would be necessary. Unfortunately, we couldn’t obtain these data for this patient population, although the thesis-generating nature of the statement seems valuable. As for ethnicity, we observed a statistically significant difference between intoxicated and other ICU patients (more Caucasian patients in the “Intoxicated” group) as well as in-between groups of different causative substances (see Table 3), but not for mortality. In general, ethnic differences in drug overdose mortality have been observed in the US in the past [18]. Thus a lower mortality among Caucasian patients, but increasing mortality among African American and Hispanic individuals was observed, yet also co-involvement of other drugs varies with ethnicity [18]. Overall, for such an analysis, a distinction must be made between prescribed and illegal opioids, for example, and co-involvement of other substances has to be investigated. Yet we cannot provide a clear explanation for our findings as an analysis for causality is too complex and beyond the scope of this study. In general, age-group stratified, socioeconomic, educational and media-triggered factors must be considered as well [18]. We also found a marked difference for initial serum lactate in survivors compared to non-survivors. This finding has been described in the past, but the proportion of deceased patients in our cohort is too small to calculate optimal cut-offs and risk groups [19]. Especially in the group of patients intoxicated with alcohols, a high initial lactate was observed. This can be explained by an altered mitochondrial metabolism with reduced utilization in both acute and chronic alcoholism [20]. However, this group showed the best outcome, which highlights the importance of initial serum lactate in the other groups. With regard to further laboratory parameters, relevant differences were shown for all blood count parameters. The lowest hemoglobin values were found in the drug toxicity group, as well as a higher median serum creatinine. A possible explanation could be the significantly higher age of these patients. The lowest platelet counts were found in the “alcohols group”. Harmful alcohol consumption is often associated with qualitative and quantitative disturbances of platelet integrity [21]. Another interesting finding is that in two groups (“drug toxicity” and “street drugs”) significantly higher leukocyte counts were observed. In general, the frequent occurrence of idiopathic leukocytosis after use of stimulant drugs (so called “uppers”) has been described in the past, especially for amphetamines [22]. With regard to higher leukocyte counts in the “drug toxicity” group, a causal explanation is again not possible in the absence of precise data regarding the causative substances. In general, a variety of drugs can cause leukocytosis, whereas a sole delimitation of infectious causes by leukocytosis is not possible in the absence of other laboratory parameters [23]. It is also interesting to note that patients with “drug toxicity” had the highest proportion of vasopressor use and dialysis in the first three days, but a zero percent short term mortality. In contrast, significantly more mechanically ventilated patients were found in the "sedatives" and "street drugs" subgroups, with these individuals again contributing the highest numerical proportion of non-survivors. This possibly underlines the need for reversibility of intoxication, as mechanical ventilation per se is a known and relatively invasive ICU measure and independent predictor of ICU mortality [24].
