In this study, we performed unpaired (between subjects) and paired (within subjects) analysis and evaluated common neuro-ICU medication use on a day-by-day basis in relation to DCI. We hypothesized that any association between medications and DCI would be a proxy measure of HA pain severity as these analgesics are largely dosed based on reported pain, however HA severity itself was not associated with DCI. Of the commonly administered medications included in the analysis, A/B/C showed an association with decreased DCI, including in day-by-day analysis where patients experiencing DCI received lower doses of A/B/C specifically on days in which they accrued DCI. In the only other previous data on a possible relationship of A/B/C and DCI, in contrary to our findings, a nearly significant association for higher doses of A/B/C with radiographic vasospasm in the first 7 days post aSAH was found (5.5 tabs/day for vasospasm vs. 4.3 tabs/day without, p = 0.055). This study also noted a nearly significant association with vasospasm and higher opioid doses in the same timespan (8.2 mg/day OME for vasospasm vs. 5.4 mg/day OME without, p = 0.056) [8]. Important differences between that study and our own include the differing timespan (7 days vs 14 days post-SAH), cohort of patients (our study excluded patients unable to verbalize pain scale scores and hence higher Hunt and Hess grades), and outcome measure (radiographic vasospasm vs. clinical DCI). The oppositional results also raise the question whether there might be different phenotypes who respond differently to caffeine/butalbital, and the consideration of caffeine’s vasoactive effect [21]—possibly leading to radiographically detectable vasospasm—but also increased blood pressure, potentially contributing to mitigation of DCI by which caffeine could appear to increase radiographic vasospasm while decreasing DCI incidence. Previous studies have shown that A/B/C and opioids are among the most commonly used analgesics among SAH patients, as reported in an international survey of headache management post-SAH and also in our previous data where ~ 90% of all post-SAH patients over a 7-year span received these analgesics while admitted [22]. Thus, it is important to clarify any association between these medications and DCI. Given the clinical equipoise resulting from the prior study and our own on these ubiquitous medications that nearly all aSAH patients receive, further study is needed to clarify whether there is an association between A/B/C and DCI, and whether there might even be a protective effect of A/B/C in DCI, with a prospective trial design examining DCI incidence being ideal.
Possible mechanisms, in addition to the mentioned effect of caffeine as mentioned above, by which this association of A/B/C with lower rates of DCI could be explained may be combinations of analgesia decreasing sympathetic vasoconstriction, butalbital alleviating cortical spreading depolarizations (CSDs) which are postulated to play a role in DCI pathogenesis, and other mechanisms such as modulation of calcitonin gene related peptides [23, 24].
The database evaluated herein was optimal for evaluating clinical risk factors in DCI occurrence as patients were not frequently sedated due to the need to assess their pain levels, and thus their neurological exams were readily evaluable for decline indicative of DCI. Concerning HA pain, we found that neither maximum nor mean daily HA pain severity scores were associated with DCI occurrence both comparing patients who experienced DCI vs not (between subjects) and between days on which individual patients experienced DCI vs not (within subjects). Prior studies by Swope et al. and Hong et al. [8, 25] found worse HA pain to be associated with vasospasm or DCI, while studies by Huckhagel et al. and Glisic et al. did not [11, 26]; of note, these studies have different outcome measures (vasospasm vs DCI), time spans examined, and inclusion criteria when compared to our study, however from these studies and our own observations there does not appear to be a strong association between HA pain severity and occurrence of DCI. Concerning other clinical characteristics, we found that many previously evaluated risk factors for vasospasm applied to DCI in this cohort, specifically there was increased prevalence of DCI in patients receiving clipping for their aneurysm treatment as opposed to coiling [27]. The increased cerebral insult from invasive surgery may contribute to the varied pathophysiologic pathways leading to DCI. Likewise, there was a noted trend (although not statistically significant) towards younger age and higher modified Fisher scores in DCI patients which has been previously reported on [1]. There was no association between aneurysm radiographic size, EVD placement, or Hunt & Hess score with DCI in this study. Uniquely, in this cohort there was an increased prevalence of DCI in men compared to women, which is at odds with literature in which DCI is more prevalent in women post aSAH [1]. The reasons for this may be related to this specific cohort of patients in which the patients with the worst Hunt & Hess grades were excluded, which is a limitation of the study in this context as it has been reported that women may experience worse symptoms with DCI and thus would be over-represented in our exclusion criteria [28]. The morbidity burden of DCI is reflected in the prolonged length of stay for DCI patients found in this cohort.
A key limitation of this study is the exclusion of more severe Hunt & Hess grade patients who would not have been able to request analgesic medications due to their clinical condition. This limitation may limit generalizability of these findings in aSAH patients with more severe disease. On the other hand – this is a strength when considering the assessment of pain severity and its relation to medication administration and DCI. Additionally, it is recognized that the dosages of analgesics received is highly correlated with the amount of headache pain the patients are experiencing, however we did not find a relationship between headache severity and DCI probability and thus headache pain does not appear to be a confounding variable. Overall, our results indicate that A/B/C, which is already used to manage headaches after an SAH, may be associated with decreased DCI incidence. Other commonly used medications in SAH patients including acetaminophen, opioids, dexamethasone, and levetiracetam were not associated with DCI. Given the widespread usage of these medications during the time window in which DCI frequently occurs, the present study aimed to investigate if these drugs have an impact on DCI incidence, which is not currently a consideration in deciding whether to prescribe these medications and at what dosages. Future studies could aim to investigate the impact of A/B/C use on DCI incidence in a prospective manner, and in a wider range of aSAH patients.