In this study we have showed that there was no clear trend in reduction of mortality rates over the years and that the occurrence of UO remained high and unchanged in poor grade SAH patients. WFNS 5 and the absence of hydrocephalus were independent predictors of hospital mortality. WFNS 5 was also independently associated with UO. The presence of intracerebral hematoma and hydrocephalus did not significantly influence the high mortality and UO rates in WFNS 5 patients.
Several studies have reported worldwide a decrease in SAH mortality over time, regardless of the outcome used (i.e., at hospital discharge or at 30 days). Indeed, two meta-analysis including studies from Europe, Japan, China, USA, Chile, Australia and New Zealand have reported a decrease in mortality of approximately 0.8 % and 0.9% per year and from 1972 to 2002 and from 1980–2005, respectively [28, 29]. In Australia, two retrospective studies have shown a decrease in mortality varying from 0.7–2.7% a year after SAH over time [30, 31]. Similarly, mortality of SAH patients has decreased steadily between 1998 and 2007, regardless of the initial treatment of choice (i.e. surgical clipping or endovascular coiling) [32]. In our study, we found a significant reduction in mortality from 2004–2007 to 2008–2011; however, mortality increased in 2012–2015 with similar rates to 2004–2007. The increase in mortality in the third period could be explained by the significant increase in the proportions of WFNS 5 compared to the second period and due to the high proportions of patients with intracranial hypertension. In fact, similar incidences of intracranial hypertension and subsequently mortality rates were observed in the first and third period. In the last studied period (2016–2018) mortality starts to fall again, despite the elevated proportions of WFNS 5 patients. This could be due to the implementation of new neuromonitoring technologies, such as PbtO2 catheters, or more aggressive therapies (i.e. intra-arterial vasodilators), which may have contributed to early diagnosis of neurological complications and a more adequate management of refractory cerebral vasospasm. This difference from what was reported in previous studies could be explained because these studies have frequently compared 80’s and 90’s with more recent decades, while we focused our data analysis in 2004–2018, where endovascular and surgical treatment were already well established; also, while previous studies often focused on SAH of all clinical severity on admission [32], we evaluated only poor grade patients, in whom the mortality rate is the highest and relevant improvement in patients’ management might significantly influence patients’ outcome over time. Of course, not all confounders could be adequately assessed in this retrospective longitudinal analysis; moreover, other series of SAH patients have reported unchanged mortality rates over the same period of time [33, 34].
Regarding functional neurological status, a study conducted in the UK including all WFNS grades showed a 50% decrease in the occurrence of UO, which was assessed by the modified Rankin scale, comparing two distinct time periods (i.e. 1981–1986 vs. 2002–2008) [28]. In the USA, a decrease in the proportion of patients with high disability was observed from 1998 to 2013 [35]. Similarly, in developing countries such as India, there was a significant decrease of patients with UO at 3 months, which was assessed by GOS, from 1996–2015; one of the main explanations for such findings was the improvement of neurosurgical services and overall therapies in this country [35]. However, when only patients with poor grade clinical status were considered, early interventions and aggressive treatment did not significantly reduce the high number of patients who were discharged with severe disability from the hospital [7, 36, 37]. A meta-analysis also showed that in poor grade SAH patients there was an initial improvement of outcome form 70’s to 90’ but that there has been no further gain in terms of neurological recovery thereafter; this could be explained by the high proportion of WFNS 5 patients included in the studies [20]. Our study reported similar results and underlined a very high occurrence of UO in WFNS 5 patients.
Our findings suggest that patients with poor grade SAH may not be an homogenous group; in particular, patients with WFNS 5 have a worse outcome than those with WFNS 4. This has been previously suggested in other studies [7, 20, 38], but our study was designed to specifically investigate the differences in outcome between WFNS 4 and 5. The highest mortality and UO rates observed in WFNS 5 patients might be due to several factors, including the extension of the initial injury, the severity of bleeding as well as the occurrence of early (i.e., intracranial hypertension) and late (i.e., DCI) brain complications. Future trials should focus on the pathophysiological mechanisms as the response to therapies of WFNS 4 and WFNS 5 patients separately to optimize therapeutic interventions in such patients, better stratify for the severity of disease and more accurately prognosticate their outcome.
We also investigated additional factors that could further aggravate or influence the poor outcome of WFNS 5 patients, such as hydrocephalus and the presence of intraparenchymal hematoma. We found that acute hydrocephalus did not increase the chance of death, possibly because clinical deterioration associated with hydrocephalus could promptly be treated with external ventricular derivation, which in many cases may resolve symptoms [39]; as such, the importance of early recognition and treatment of this condition should be highlighted and reported in future descriptive and interventional studies. As another study reported that hydrocephalus after SAH was associated with UO [40], this complication should be further studied in multicentric cohorts. In our study, we also reported a higher prevalence of intraparenchymal hematoma than in previous studies [22]. However, this had no impact on mortality or UO. A possible explanation is that hematoma evacuation surgery was performed in 82% of patients and this might have reduced the risk of secondary brain injury [22]. Also, it is possible that the association of intraparenchymal hematoma with UO reported in other studies might be related to the decision of limiting aggressive therapies, including surgical drainage.
Potential limitations of the present study should be taken consideration. First, many additional variables could not be collected as data availability and quality can be challenging in retrospective studies covering a long period of time. Also, decision of specific therapies could have been influenced by the patient status, physician or family decisions and not solely on patient’s severity which may have affected outcome. Finally, the follow-up period for neurological outcome could have been prolonged to 6–12 months to better investigate the long-term evolution of such poor grade SAH patients.