Preventable poor outcome from rebleeding by emergency aneurysm occlusion in patients with aneurysmal subarachnoid haemorrhage

Abstract

Introduction

The risk of rebleeding is highest during the initial hours after aneurysmal subarachnoid haemorrhage (aSAH), but the aneurysm is not occluded in all patients immediately after admission.

Our aim was to determine the proportion of aSAH patients with poor outcome from early in-hospital rebleeding that can be prevented by three emergency aneurysm occlusion regimes.

Patients and methods

From our prospectively collected database, we retrieved from all aSAH patients admitted between July 2007 and July 2017 data on clinical condition on admission, time of rebleeding, and outcome at 3 months.

Results

Of 1391 consecutive aSAH patients, 923 were in good clinical condition and had an aneurysm on initial imaging that was amenable for treatment. Poor outcome from rebleeding could have been avoided by treatment <4 h during day time shifts in 4 (0.4% [95% CI: 0.2–1.1]) patients (number needed to treat [NNT]: 250), by treatment and <1 h during daytime shift in 9 (1.0% [95% CI: 0.5–1.8]; NNT: 111), and treatment <1 h at 24/7 basis in 16 (1.7% [95% CI: 1.1–2.8%]; NNT: 59).

Discussion

Emergency aneurysm occlusion can reduce poor outcome due to rebleeding, but only in small proportions of patients. Whether such strategies lead to improved outcome for all patients and are cost-effective is highly uncertain.

Conclusion

We do not recommend instalment of a treatment regimen where occlusion of ruptured aneurysm is performed within 1 h on a 24/7 basis.

Introduction

In patients with aneurysmal subarachnoid haemorrhage (aSAH), rebleeding is still a major cause of poor outcome.¹ The incidence of rebleeding is the highest on the day of the initial haemorrhage, and in particular, within the first hours.^2–4 Intervention within 72 h, and in particular within 24 h, has been shown to reduce the proportion of patients with rebleeding,^5–7 but not all patients receive aneurysm occlusion within one or a few hours after admission. In some patients, treatment is deferred because of a poor clinical condition on admission, in others because of logistic reasons, or because of the lack of availability of a 24/7 team to treat early also in out-of-office hours. Accordingly, in-hospital rebleeding still occurs in patients with aSAH.

Previous studies on in-hospital rebleeding focused on a reduction of all episodes of rebleeding. However, the impact on outcome is not similar for all episodes of rebleeding. The impact of rebleeding is much higher in a patient who is in a good clinical condition waiting for aneurysm occlusion than in a patient who is already in a poor condition. Also, patients who are in a good clinical condition on admission can have good outcome despite rebleeding.

Since the goal of treating aSAH patients is not prevention of rebleeding, but increasing the chance of a good outcome, we studied, within a setting aiming to occlude ruptured aneurysms <24 h after admission, the proportion of patients with aSAH who were admitted in good clinical condition and had a poor condition after rebleeding. We assessed the proportion of poor outcome from rebleeding that could have been prevented by aneurysm occlusion within 4 h during daytime shift, within 1 h during daytime shift or within 1 h on a 24/7 basis.

Methods

We retrieved data from consecutive patients, admitted in good clinical condition to our tertiary referral centre (University Medical Center Utrecht [UMCU], Utrecht, The Netherlands) between 1 July 2007 and 1 July 2017 with aSAH. The diagnosis of aSAH was based on the presence of blood on plain head CT or the presence of bilirubin in cerebrospinal fluid at least 12 h after ictus, in combination with the presence of an aneurysm on CT angiography (CTA), digital subtraction angiography (DSA) or magnetic resonance angiography (MRA).

During the study period, the hospital policy was to perform aneurysm occlusion during the first day time or evening shift if patients were admitted in good clinical condition. In patients admitted in poor clinical condition, clinical improvement was awaited for approximately 72 h, except in case of associated intracerebral or subdural haematoma necessitating emergency evacuation, in which instances the aneurysm was occluded in the same procedure. Also, in patients in poor clinical condition needing extra-ventricular drainage, clinical improvement after drainage was awaited before aneurysm treatment. Despite this policy, logistic reasons could have resulted in postponed treatment.

Events were defined as episodes of rebleeding occurring between 1 and 48 h after admission. Episodes of rebleeding less than 1 h after admission were not considered an event, since we deemed aneurysm occlusion within less than 1 h not feasible. Rebleeding was defined as a sudden decrease in the level of consciousness, with or without new focal signs, with a head CT showing an increase of subarachnoid, intracerebral, or intraventricular blood in comparison with the preceding scan.

The day of admission was calculated from the day of SAH (= day 0). The time of rebleeding in hours after admission was defined as the time interval between the time of the head CT scan revealing the rebleeding and the time of presentation on the emergency department. We extracted data on clinical condition on admission and shortly before rebleeding, location of the aneurysm and non-logistic reasons for postponing aneurysm occlusion.

The clinical condition was categorised according to the Prognosis on Admission of Aneurysmal Subarachnoid Hemorrhage (PAASH) score.⁸ A good condition was defined as a PAASH grade I–III, and a poor condition as PAASH grade IV or V. Functional outcome was assessed with the Glasgow Outcome Scale (GOS) at 3 months after ictus.⁹ Poor functional outcome was defined as GOS 1–3, and good functional outcome as GOS 4–5. Poor outcome from rebleeding was defined as a patient in a good clinical condition before rebleeding and a poor outcome after rebleeding. Poor outcome from rebleeding was considered preventable if a patient was admitted in good clinical condition and had rebleeding between 1 and 48 h after admission, while there was no diagnostic delay or a clinical reason for postponing treatment.

Statistical analysis

We used descriptive statistics to describe characteristics of our study population and calculated proportions with corresponding 95% confidence intervals (CI) of patients in whom a poor outcome from rebleeding could have been prevented by aneurysm occlusion. For this purpose, we calculated proportions for three different scenarios: (1) aneurysm occlusion within 4 h in the first daytime shift; (2) within 1 h in the first daytime shift; or (3) within 1 h on a 24/7 basis.

Ethical aspects

The institutional research ethics of board of UMCU decided that no formal assessment was needed because of the retrospective design of the study and because all patient data were anonymised.

Results

Clinical characteristics and timing of transfer or admission

Of the 1391 patients with aneurysmal SAH admitted from 1 July 2007 through 1 July 2017, 1035 were admitted in a good clinical condition (Figure 1). Initial angiography showed an aneurysm in 938 of these patients. Of patients with rebleeding, 15 had clinical reasons for postponing intervention, therefore rebleeding was not considered preventable by emergency aneurysm closure. Thus, 923 patients were admitted in good clinical condition with no reasons for postponing intervention (Figure 1 and Table 1). In-hospital rebleeding occurred in 25/923 (2.7% [95% CI 1.8–4.0]) patients before intervention. In four of these patients rebleeding had occurred within 1 h after admission and was thus considered not-preventable by emergency occlusion. In total, 21/923 (2.3% [95% CI: 1.5–3.5]) patients admitted in good clinical condition had in-hospital rebleeding between 1 and 48 h after admission. Poor outcome occurred in 16 (1.7% [95% CI: 1.1–2.8]) of these patients (Table 2).

Figure 1.

Flowchart of patient selection.

Table 1.

Baseline characteristics of 923 aSAH patients admitted in good clinical condition (PAASH I–III) with no clinical reason for postponing intervention.

Characteristics	No in-hospital rebleedingn (%)	In-hospital rebleedingn (%)
Patients	898	25
Women	656 (73)	18 (72)
Median age (range)	57 (18–92)	62 (29–86)
Admission to our hospital after SAH
Day 0	512 (57)	18 (72)
Day 1	190 (21)	4 (16)
Day 2 or later	196 (22)	3 (12)
Time of admission
Night (0:00–6:00 h)	161 (18)	5 (20)
Morning (6:00–12:00 h)	120 (13)	2 (8)
Afternoon (12:00–18:00 h)	329 (37)	12 (48)
Evening (18:00–24:00 h)	288 (32)	6 (24)
PAASH score at admission
I (GCS 15)	460 (51)	7 (28)
II (GCS 11–14)	344 (38)	13 (52)
III (GCS 8–10)	94 (11)	5 (20)
Aneurysm location
Anterior circulation	569 (61)	17 (68)
Posterior circulation	329 (38)	8 (32)

aSAH: aneurysmal subarachnoid haemorrhage; PAASH: Prognosis on Admission of Aneurysmal Subarachnoid Haemorrhage scale; GCS: Glasgow Coma Scale.

Table 2.

Clinical condition at time of rebleeding and GOS at 3 months for 21 patients with rebleeding 1–48 h after admission.

Neurological condition before rebleeding according PAASH scale	n	Glasgow Outcome Scale score at discharge
		5	4	3	2	1
I (GCS 15)	6	–	–	3	–	3
II (GCS 14–11)	12	1	3	2	–	6
III (GCS 8–10)	3	5	–	–	–	2

GOS: Glasgow Outcome Scale, 5: excellent, 4: good; 3: fair; 2: poor; 1: dead.⁹

Time of rebleeding and potential prevention

Table 3 shows details the timing of rebleeding, and Figure 2 shows the time interval between admission and rebleeding for the 16 patients admitted in good condition with poor outcome after in-hospital rebleeding. Poor outcome from rebleeding could have been prevented in 4 patients (0.4% [95% CI: 0.2–1.1]) if treatment in all patients was performed <4 h in daytime shift after admission, in 9 patients (1.0% [95% CI: 0.5–1.8]) if treatment was performed within 1 h during daytime, and in 16 patients (1.7% [95% CI 1.1 – 2.8]) if treatment was performed within 1 hour on a 24/7 basis. The corresponding numbers need to treat to prevent one episode of poor outcome from rebleeding are 250 for treatment <4 h in daytime shift, 111 for treatment <1 h in daytime shift and 59 for treatment <1 h on a 24/7 basis.

Table 3.

Time of rebleeding of the 21 patients who had a rebleeding 1–48 h after admission.

Hours after admission	h:min
Median hours (range)	3 h: 18 min (1 h: 07 min–32 h: 08 min)
Mean	5 h: 45 min
Circadian distribution of rebleeding	n (%)
Night (00:00–06:00 h)	6 (30)
Morning (06:00–12:00 h)	1 (5)
Afternoon (12:00–18:00 h)	7 (33)
Evening (18:00–24:00 h)	7 (33)

Figure 2.

Timing of admission and rebleeding in 16 patients with PAASH. I–III before rebleeding who died or became severely disabled. The starting point of the arrow represents the time of admission and the head of the arrow represents the time of rebleeding. Office hours are shown in white and time outside office hours are shown in grey.

Discussion

In a centre aiming for aneurysm occlusion <24 h after admission in patients with aSAH, the number of patients needed to undergo immediate treatment to further decrease the proportion of patients with poor outcome from rebleeding are large, ranging from 250 for treatment <4 h during daytime to 59 for treating all patients within 1 h after admission.

The incidence and timing of in-hospital rebleeding in our study is in line with incidences observed in previous studies, which ranged between 5.6% and 12%.^7,10–13 Two previous studies investigated the proportion of patients with rebleeding resulting in death or disability, that could have been prevented with an emergency treatment protocol. A study from South Korea demonstrated that 2/100 (2.0%) episodes of rebleeding could have been prevented with a protocol for emergency aneurysm treatment (3.5 h mean time from admission to intervention).⁷ However, since that study used historical controls, other factors than change in timing of intervention may have contributed to the improved outcome in the later study period wherein treatment was performed earlier. In contrast, a study from Poland reported that only 0.3% of all episodes of rebleedings in aSAH patients could have theoretically been prevented when immediate treatment was performed.¹² A possible explanation for this small percentage might be the distinctly low overall in-hospital rebleeding incidence (3.2%). Both these studies used proportions of all aSAH patients, whereas we only focused on those with a good clinical condition on admission, without diagnostic delay or other reasons for postponing treatment. Furthermore, the other two studies included all instances of rebleeding, whereas we only included those with poor outcome after rebleeding.

Although the European Stroke Organization guidelines recommend aneurysm treatment <72 h after ictus,¹⁴ we chose the inclusion of patients with rebleeding within 48 h after admission, because treatment beyond 48 h after ictus is mostly due to specific clinical reasons, such as difficult aneurysm configuration, signs of cerebral ischaemia or negative initial imaging. We considered a separate analysis of patients with rebleeding within 72 h after ictus to be irrelevant.

This study has limitations. First, because this is retrospective study, it is subject to inherent bias. Second, this was a monocentre study and therefore the results may have limited generalisability. Third, we considered rebleeding to be preventable if patients were admitted in good clinical condition, because admission in poor condition is a reason to postpone treatment. However, by doing this we excluded those patients who were in poor condition on admission and recovered to moderate to good conditions, thereby possibly underestimating the number of patients that could have benefited from emergency aneurysm repair. Fourth, in our institutional regimen we never occlude aneurysms in patients in good clinical condition <1 h after admission. However, we aim to occlude <24 h in patients with good clinical condition, and a proportion of patients in this series has undergone aneurysm occlusion <4 h in day time shifts. Thus, the proportion of poor outcome from rebleeding that can be prevented with this regime probably is an underestimation. Next, we did not report on any periprocedural complications. An emergency treatment regimen might be accompanied with suboptimal conditions and, consequently, increased risk of complications. Finally, in patients admitted in good condition and left in poor condition after rebleeding, we assume the poor outcome is due to rebleeding itself. This is not always the case, for outcome is influenced by multiple factors and can be a result of complications other than rebleeding.

Although emergency aneurysm occlusion can reduce the number of patients with poor outcome due to rebleeding,^6,7 it is uncertain whether such a treatment regimen should be implemented. First, the proportion of patients in whom poor outcome from rebleeding is prevented is small. Second, while some recent studies suggest ultra-early aneurysm repair is associated with better outcomes at 3 months,^5,10 one large observational study failed to show this correlation, concluding that aneurysm treatment within 24 h was not associated with better outcomes compared to treatment 24–72 h after ictus and might even be detrimental.¹⁵ Moreover, emergency occlusion, if applied to all patients, carries risks such as an increase in periprocedural complications or the induction of delayed cerebral ischaemia (DCI).² Third, emergency aneurysm occlusion implies a high logistic burden and therefore high costs. A formal cost-effectiveness assessment is outside the scope of the current study, but having a team consisting of at least vascular neurosurgeons, interventional neuroradiologists, and neuro-anaesthesiologists available at a 24/7 basis for reducing slightly the number of patients with rebleeding leading to poor outcome may not be cost-effective. Preferably, the viability of our proposed regimen would be studied in a randomised setting, taking into account the cost-effectiveness and risks. However, given the small percentage of salvageable patients and limited effect of emergency treatment, the size of such a trial would exceed 2500 patients, making it unfeasible. Meanwhile, other ways to prevent early rebleeding, such as earlier referral or the early administration of tranexamic acid should be studied. One such study is currently underway.¹⁶

Conclusion

Our results indicate that emergency aneurysm occlusion can reduce the number of patients with poor outcome due to rebleeding. We do not recommend instalment of a treatment regimen where occlusion of ruptured aneurysm is performed within 1 h on a 24/7 basis, given uncertainty of effectiveness on poor outcome overall and high costs.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

The institutional research ethics of board of UMCU decided that no formal assessment was needed because of the retrospective design of the study and because all patient data were anonymised.

Guarantor

AEL.

Informed consent

Patients were not contacted, as this is a registry-based study and thus patients consent was not obtained.

Contributorship

AEL, MW and GR researched literature and conceived the study. AEL supervised the study. RWPT collected data, conducted statistical analysis and wrote the first draft of the manuscript. All authors reviewed and edited the manuscript for important intellectual content and approved the final version of the manuscript.