In this observational cohort study of 3604 patients who were admitted to the ICU, the current findings indicate that there was an association between the use of statins and the occurrence of delirium but only in patients with mild to moderate disease severity. Specifically, the delirium occurrence was higher in the statin group with an APACHE-II score of 10-20. These findings were obtained regardless of the type of statin administered (lipophilic or hydrophilic), both before and after propensity score weighting.
To our knowledge, this is the first study to analyze how the association between the use of statins and the occurrence of delirium varied according to the disease severity. The results of these analyses suggest that the use of statins may increase the risk of delirium occurrence in patients with mild to moderate severity. This contradicts, to a degree, the existing hypothesis that the pleiotropic effects of statins, which include anti-inflammation and immunomodulation, may prevent delirium.
It can be argued that the specific underlying condition for which statins are prescribed may increase the risk of delirium in patients with mild to moderate disease severity, and not the statins themselves. However, some meta-analysis and cohort studies demonstrated that some of the diseases for which statins are recommended according to the guidelines, such as atherosclerotic cardiovascular disease and dyslipidemia [14], did not seem to be associated with delirium, or that delirium and such diseases are mutually exclusive [38–40]. Nonetheless, the effect of disease severity on the association between the specific underlying disease and delirium is still unknown. Thus, the hypothesis that the underlying conditions requiring statin treatment may have influenced our results is not supported by previous evidence. In this study, we discuss the effect of statin and disease severity on biological mechanisms, which are more specific than underlying disease conditions, such as the role of nitric oxide (NO) in the systemic inflammation state and hypoperfusion of the brain.
In the inflammatory state, cytokines trigger over-production of inducible nitric oxide synthase (iNOS), which is inactive under normal physiological conditions. This is known as one of the crucial mechanisms of sepsis, by which large amounts of NO are produced [41]. Overproduction of NO leads to changes in the vascular tone by affecting the endothelium and small blood vessels involved in systemic circulation, and causes hypotension and migration of leukocytes, which results in widespread tissue damage and multiple organ failure [41, 42]. Statin has been shown to lead to concomitant up-regulation of endothelial nitric oxide synthase (eNOS), which is an isoform of iNOS, and down-regulation of iNOS. As such, statin modulates the production of NO and restores systemic circulation. Consequently, statin may help in attenuating inflammation and even delirium severity [18, 43].
However, some researchers suggested that during inflammation, the modulation of NO by statins may impact cerebral autoregulation resulting in cerebral ischemia [44, 45]. When the microvascular tone is recovered by the modulation of NO, the blood flow is distributed to peripheral small blood vessels, which reduces the blood flow to the brain [24, 25]. Hypoperfusion and inadequate cerebral oxygenation play an important role in the pathophysiology of delirium [10, 46, 47]. If this hypothesis applies to our results, it can be inferred that patients with mild to moderate disease severity under statin treatment may be more susceptible to the delirium-inducing effect of statins through hypoperfusion than the delirium-preventing effect through anti-inflammation.
As for the statin type, use of both lipophilic and hydrophilic statins in patients with mild to moderate severity was associated with delirium, but there were no significant differences between the two. Hydrophilic statins have been reported to up-regulate eNOS expression to levels comparable to those seen with lipophilic statins [43, 48], but there is still some uncertainty about how much the overall expression of NO differs depending on the statin types. Our results suggest that the NO synthesis may not significantly vary according to the statin types.
In contrast, except for group 2, the use of statins in the remaining groups was not significantly associated with the occurrence of delirium, regardless of the statin types. Since patients in group 1 were in a relatively less inflammatory state than those in group 2, we assume that the amount of iNOS was much lower in group 1. Thus, the balancing effect of statins on NO synthases in peripheral blood vessels and the accompanying hypoperfusion effect on the brain may have also been insignificant. Meanwhile, in groups 3 and 4, the higher disease severity may have played the primary role in the occurrence of delirium [10]; however, the function of statins in these groups remaining unclear.
This study has several limitations. First, due to the multifactorial nature of delirium, the potential risk factors which may have affected the outcomes were not sufficiently controlled, except for age and disease severity. Second, the differential roles of pitavastatin and pravastatin were difficult to examine in either statin group, because both of them were under-represented; the lipophilic statin group consisted mostly of atorvastatin treatment while the other group was mainly represented by rosuvastatin. Lastly, it is difficult to generalize our results because this study included data from a single hospital.