Association of dexamethasone with shunt requirement, early disability, and medical complications in aneurysmal subarachnoid hemorrhage

Abstract

BACKGROUND AND PURPOSE:

Current guidelines do not support the routine use of corticosteroids in patients with aneurysmal subarachnoid hemorrhage (aSAH). However, corticosteroids use in aSAH has been practiced at some centers by convention. The aim of the study was to determine the incidence of hydrocephalus requiring ventriculoperitoneal shunt (VPS) placement as well as functional outcome on discharge and adverse events attributed to corticosteroids in patients with aSAH treated with different dexamethasone (DXM) treatment schemes.

METHODS:

We retrospectively analyzed 206 patients with aSAH stratified to three groups based on the DXM treatment scheme: no corticosteroids, short course of DXM (S-DXM; 4mg every 6 hours for 1 day followed by a daily total dose reduction by 25% and then by 50% on last day), and long course of DXM (L-DXM; 4mg every 6 hours for 5–7 days followed by reduction by 50% every other day). The primary outcome measure was the placement of a VPS and the secondary outcome was a good functional outcome (modified Rankin Scale [mRS] 0–3) at hospital discharge. Safety measures were the incidence of infection (pneumonia, urinary tract infection, ventriculitis, meningitis), presence of delirium, and hyperglycemia.

RESULTS:

There was no difference in the rate of external ventricular drain (EVD) (p=0.164) and VPS placement (p=0.792), nor in the rate of good outcome (p=0.928) among three defined treatment regimens. Moreover, the median duration of treatment with EVD did not differ between subjects treated with no corticosteroids, S-DMX, and L-DXM (p=0.905) and the probability of EVD removal was similar when stratified according to treatment regimens (log-rank; p=0.256). Patients who received L-DXM had significantly more complications as compared to patients, who received no corticosteroids or S-DXM (78.4% vs. 58.6%; p=0.005). After adjustment, L-DXM remained independently associated with increased risk of combined adverse events (OR=2.72; 95%CI, 1.30–5.27; p=0.008), infection (OR=3.45; 95%CI, 1.63–7.30; p=0.001) and hyperglycemia (OR=2.05; 95%CI, 1.04–4.04; p=0.039).

CONCLUSIONS:

DXM use among patients with aSAH did not relate to the rate of EVD and VPS placement, duration of EVD treatment, and functional disability at discharge but increased the risk of medical complications.

Introduction

Hydrocephalus, a complication of aneurysmal subarachnoid hemorrhage (aSAH), is associated with worse neurologic outcomes as well as overall increased morbidity and mortality^1–4. The reported incidence of symptomatic hydrocephalus requiring external ventricular drain (EVD) placement ranges from 7% to 65%⁵ and approximately one third of patients require chronic cerebrospinal fluid (CSF) diversion with ventriculoperitoneal shunt (VPS)^1,6–9.

Currently corticosteroid treatment in aSAH remains controversial^10–12. A small randomized, double-blind, placebo-controlled trial showed potential efficacy of corticosteroids (3-day treatment of methylprednisolone 16mg/kg) by improving functional outcomes one year after aSAH as compared with placebo¹³; however, the results were limited by the modest sample size (n=94). To our knowledge there have been no other randomized clinical trials conducted to elucidate this issue. Given the limited data, current guidelines do not support the routine use of corticosteroids in patients with aSAH^14,15 and a number of concerns have been raised that treatment with high dose corticosteroids increase the risk for adverse events including infections, hyperglycemia, and delirium.

In this study, we sought to determine the incidence of hydrocephalus requiring VPS placement as well as the functional outcome on discharge assessed by modified Rankin Scale (mRS) in patients with aSAH treated without dexamethasone (DXM) versus short and long course of DXM. In addition, we stratified analyses according to different DXM regimens to gain potential insight into length of treatment dependent effects.

Methods

Study cohort

Using local stroke registries, we retrospectively identified adult patients with aSAH that were consecutively hospitalized at the University of Massachusetts Memorial Medical Center (UMMMC), in Worcester, MA from January 2009 to December 2014 and Rhode Island Hospital (RIH), in Providence, RI from January 2015 to January 2017. Institutional Review Board approval was obtained from each of the participating centers.

A priori defined exclusion criteria were age <18 years, non-aneurysmal SAH, exam consistent with brain death on initial presentation, transition to comfort measures only within 24 hours after admission, pregnancy and diabetes mellitus type 1. Furthermore, we excluded subjects, who were administered corticosteroids other than DXM (i.e. methylprednisolone, hydrocortisone) as well as patients who received a DXM regimen that differed from the short and long course of DXM regimens as defined for the purpose of this study (see below for details).

Age, gender, presence of hypertension (defined as use of antihypertensive medications, or the diagnosis made in the outpatient setting based on systolic blood pressure of ≥140 mm Hg or diastolic blood pressure of ≥90 mm Hg on 2 separate occasions), smoking status, EVD placement, treatment method (surgical clipping, endovascular intervention, conservative management) were collected for each subject. The severity of aSAH on admission was assessed according to Glasgow Coma Scale (GCS), Hunt and Hess (HH), and Word Federation of Neurosurgical Societies (WFNS) grade. The highest GCS score documented within the first 24 hours after admission was included in the analysis.

All computed tomography (CT), CT angiography scans, and cerebral angiograms were reviewed to determine aneurysm location, presence of hydrocephalus, intraventricular hemorrhage (IVH) on admission, severity of aSAH according to modified Fisher Scale (mFS), and the presence of delayed cerebral ischemia (DCI).

Treatment protocol

The studied population was stratified to three groups based on the standard DXM treatment schemes in our neurointensive care units during their index hospitalization: 1) patients who did not receive DXM (or other corticosteroids); 2) patients who received a short course of DXM (S-DXM; defined as 4mg every 6 hours for 1 day followed by a daily total dose reduction by 25% and then by 50% on last day); 3) patients who received a long course of DXM (L-DXM; defined as 4mg every 6 hours for 5–7 days followed by reduction by 50% every other day). The use of the different DXM treatment schemes was at the discretion of the attending neurosurgeon and neurointensivist. Patients receiving a corticosteroid other than the those defined above were excluded from this study. L-DXM was administered to all patients recruited at UMMMC per institutional protocol. The indications for S-DXM at RIH included post-craniectomy steroid taper and intractable headache.

Both institutions utilized gradual weaning protocols where EVD was raised by 5cm H₂O each day until 20cm H₂O was reached and then clamped and removed if patient was clinically stable, had no ICP elevations and had no radiological hydrocephalus present on CT. The decision to place VPS was based on recurrent symptomatic hydrocephalus post-removal or a failed weaning EVD trial – the development of symptomatic hydrocephalus – which was a clinical judgement call by the attending neurointensivist and neurosurgeon.

Outcome measures

The primary outcome measure of interest was the placement of a VPS. The decision regarding the placement of VPS was at the discretion of treating neurosurgeon and neurointensivist.

A secondary outcome measure was the mRS score at hospital discharge. The mRS score documented by physician or nurse practitioner certified in mRS or reconstructed based on case description, was abstracted from medical records. Outcome on discharge was dichotomized to good (mRS score 0–3) versus poor (mRS score 4–6).

The primary safety measure included the composite of infection (pneumonia, urinary tract infection, ventriculitis, meningitis), delirium, or hyperglycemia. Secondary safety measures included the individual components of the primary safety measure.

Pneumonia and urinary tract infection were diagnosed based on Centers for Disease Control and Prevention/National Healthcare Safety Network surveillance definitions¹⁶. Ventriculitis/meningitis related to EVD was defined based on modified criteria adopted for intensive care unit settings¹⁷. Hyperglycemia was defined as blood glucose value >140 mg/dl for 2 consecutive results. The presence of delirium was abstracted from medical records and diagnosed clinically and using Richmond Agitation-Sedation Scale (RASS)¹⁸.

Statistics

Unless otherwise stated, continuous variables are reported as mean ± S.D. or median (interquartile range). Categorical variables are reported as proportions. Between-group comparisons for continuous and ordinal variables were made with Mann-Whitney U test and Kruskal–Wallis one-way ANOVA on ranks as appropriate. Categorical variables were compared using the χ²-test. Bonferroni method was used to correct for multiple comparisons. Kaplan–Meier and log rank tests were used for univariate comparisons of the duration of treatment with EVD.

We created separate multivariable logistic regression models to determine factors associated with a lower VPS placement rate and a good functional outcome on discharge (dependent variables). In addition, we created separate regression models to determine factors associated with primary and secondary safety measures. Given the low number of cases with ventriculitis/meningitis (n=3) no attempt was undertaken to construct multivariable regression models to determine independent risk factors. All models were adjusted for age, gender, GCS, mFS scores and corticosteroid regimen (none, S-DXM, L-DXM).

To avoid model overfitting in the multivariable models, variables were sequentially removed (likelihood ratio) from the models at a significance level of 0.1. Collinearity diagnostics were performed (and its presence rejected) for all multivariable regression models. Model calibration was assessed by the Hosmer-Lemeshow test and model fit determined by examining the −2 log-likelihood statistic and its associated chi-square statistics.

Two-sided significance tests were used throughout and unless stated otherwise a two-sided p<0.05 was considered statistically significant. All statistical analyses were performed using IBM® SPSS® Statistics Version 22 (IBM®-Armonk, NY).

Results

Study cohort

We identified 292 patients (n=180 at UMMHC; n=112 at RIH) diagnosed with suspected aSAH. Of these, 206 patients fulfilled the study criteria and were included for analysis (Figure 1). Data was complete for all variables except for smoking status (n=5 [2.4%] missing). Table 1 summarizes the baseline characteristics of the studied population. Except for age (p=0.007) and the distribution of the treatment method (p<0.001), there was no difference in basic risk factors, admission HH, GCS, WFNS and mFS scores between subjects treated with no corticosteroids, S-DXM, and L-DXM (p>0.05, each) in unadjusted analyses. Furthermore, there was no difference in the presence of hydrocephalus, IVH, EVD placement and DCI between treatment groups (p>0.05, each). The median duration of treatment (days) with EVD did not differ between subjects treated with no corticosteroids, S-DMX, and L-DXM (p=0.905, Tab.1). Similarly, there was no difference in the mean duration of treatment with EVD when we compared subjects that received no corticosteroids or S-DXM with subjects receiving L-DXM (p=0.706) and when we compared subjects who received no corticosteroids versus the combined S-DXM and L-DXM groups (p=0.902). Among surviving patients who did not have a VPS placed, Kaplan–Meier analysis indicated similar probability of EVD removal when stratified according to treatment with no corticosteroids, S-DXM and L-DXM (log-rank p=0.256, Fig. 2). Excluding patients with EVD placement after 48 hours (n=9) from our analysis did not meaningfully change the results (not shown). In 25 (12.1%) patients, the aneurysm was located in the posterior circulation; in 124 (62.2%) patients, the aneurysm was located in the anterior circulation; and in 57 (25.7%) patients, aneurysm was found in more than one location.

Figure 1. Flow chart of study design and patient selection.

CMO=comfort measures only; DM=diabetes mellitus; L-DXM=long course of dexamethasone; HC=hydrocortisone; S-DXM=short course of dexamethasone; MTP=methylprednisone; SAH=subarachnoid hemorrhage.

Table 1.

Baseline characteristics, primary and secondary outcome and safety measures of the studied patient population as stratified by the duration of dexamethasone (DXM) treatment.

Characteristics	All patients (n=206)	No steroids (n=37)	S-DXM (n=21)	L-DXM (n=148)	p-value
Age, years	55.0 (48.8–63.7)	59.9 (54.7–68.7)	51.6 (40.8–61.3)	54.0 (48.5–63.0)	0.007
Female gender	142 (68.9)	27 (73.0)	15 (71.4)	100 (67.6)	0.790
Hypertension	102 (49.5)	22 (59.5)	8 (38.1)	72 (48.6)	0.272
Smoking	114 (55.3)	21 (56.8)	14 (66.6)	79 (53.4)	0.252
Aneurysm treatment					<0.001
Endovascular	160 (77.7)	31 (83.8)	7 (33.3)	122 (82.4)
Open surgery	28 (13.6)	0 (0)	13 (61.9)	15 (10.1)
Other*	18 (8.7)	6 (16.2)	1 (4.8)	11 (7.4)
Hydrocephalus	123 (59.7)	26 (70.3)	12 (57.1)	85 (57.4)	0.351
Presence of IVH	103 (50.0)	21 (56.8)	10 (47.6)	72 (48.6)	0.660
Hunt and Hess	3 (2–4)	3 (1–4)	3 (1–3)	3 (2–4)	0.336
GCS	14 (8–15)	14 (9–15)	14 (10–15)	13.5 (8–15)	0.695
WFNS	2 (1–4)	2 (1–4)	2 (1–4)	2 (1–4)	0.878
mFS	3 (1–4)	2 (1–3)	3 (1–3)	3 (1–4)	0.526
EVD	125 (60.7)	27 (73.0)	14 (66.7)	84 (56.8)	0.164
Duration of EVD treatment [days]	11 (6.0–15.0)	11 (5.5–16.5)	10 (6.0–17.0)	11 (6.5–15.0)	0.905
DCI	62 (30.1)	10 (27.0)	7 (33.3)	45 (30.4)	0.871
VPS	14 (6.8)	3 (8.1)	2 (9.5)	9 (6.1)	0.792
Discharge mRS 0–3	120 (53.3)	21 (56.8)	13 (61.9)	86 (58.1)	0.928
Pneumonia	62 (30.1)	5 (13.5)	4 (19.0)	53 (35.8)	0.015
Urinary tract infection	49 (23.8)	4 (10.8)	3 (14.3)	42 (28.4)	0.045
V entriculitis/meningitis	3 (1.6)	1 (2.7)	0 (0)	2 (1.4)	0.697
Hyperglycemia	122 (59.2)	19 (51.4)	8 (38.1)	95 (64.2)	0.042
Delirium	26 (12.6)	6 (16.2)	2 (9.5)	18 (12.2)	0.724

Data are n (%) or median (IQR). DCI=delayed cerebral ischemia; EVD=extraventricular drain; GCS=Glasgow Coma Scale; IVH=intraventricular hemorrhage; mFS= modified Fisher Scale; mRS=modified Rankin Scale; VPS=ventriculoperitoneal shunt; WFNS=Word Federation of Neurosurgical Societies.

^*- both, endovascular and open surgery, or conservative management.

Figure 2. Probability of treatment with EVD (Kaplan-Meier analysis).

Kaplan-Meier curves for the cumulative probability of external ventricular drain (EVD) removal stratified by treatment paradigm (blue = no corticosteroids; green = short course of dexamethasone [S-DXM]); orange = long course of dexamethasone [L-DXM]) showed no between group difference (p = 0.256).

Overall, there were no differences in the baseline clinical and radiological characteristics of patients recruited from UMMMC (n=148) versus RIH (n=58) except for the more frequent surgical approach at RIH as compared to UMMMC (Supplemental Table 1).

Treatment with DXM was not associated with the rate of VPS placement

Overall, there was no difference in the rate of VPS placement between the three defined treatment regimens (Table1). When we combined no corticosteroids with S-DXM groups and compared these with L-DXM results were similar (p=0.543). Likewise, there was no difference in VPS placement between the patients who received no corticosteroids versus any (S-DMX/L-DXM) of DXM treatment regimen (p=0.720). Neither S-DXM, L-DXM, or any DXM treatment regimen (combined S-DXM and L-DXM) were associated with the rate VPS placement on multivariable regression (p>0.005, each; Table 2). Additional layered analysis in groups defined by (1) HH grade on admission (HH 0-II and III-V), (2) the presence or absence of IVH (3) the treatment method and (4) the requirement for EVD placement on admission did not show the effect of S-DMX, L-DXM or S-DXM/L-DXM on the rate of VPS placement (p>0.05, each, not shown).

Table 2.

Multivariable logistic regression models for factors associated with the primary and secondary outcome measures.

Factors associated with the ventriculoperitoneal shunt placement (1)
	Crude OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Age (years)	0.99 (0.95–1.04)	0.849	0.99 (0.95–1.04)	0.683
Female gender	0.59 (0.16–2.18)	0.424	1.81 (0.47–6.99)	0.390
GCS score	0.94 (0.82–1.06)	0.308	0.95 (0.82–1.09)	0.458
mFS score	1.23 (0.78–1.96)	0.373	1.15 (0.67–1.96)	0.609
L-DXM	1.46 (0.47–4.55)	0.517	0.68 (0.21–2.15)	0.507
Factors associated with the incidence of good outcome at discharge (2)
	Crude OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Age (years)	0.98 (0.96–0.99)	0.040	-	-
Female gender	1.49 (0.82–2.69)	0.192	-	-
GCS score	1.57(1.37–1.79)	<0.001	1.41 (1.26–1.57)	<0.001
mFS score	0.36 (0.27–0.48)	<0.001	0.53 (0.38–0.74)	<0.001
*L-DXM	0.93 (0.50–1.71)	0.805	1.41 (0.61–3.23)	0.424

We used p<0.05 as criteria for backward steps. Blank cells with dashes represent variables that were dropped as non-significant during stepwise selection.

⁽¹⁾Hosmer-Lemeshow goodness of fit χ² 9.06, p=0.337.

⁽²⁾Hosmer-Lemeshow goodness of fit χ² 6.04, p=0.535. GCS=Glasgow Coma Scale; L-DXM=long course of dexamethasone; mFS=modified Fisher Scale.

^*No steroids and S-DXM (combined) were used as reference group in the model presented in the table.

Treatment with DXM was not associated with functional disability at discharge

We found no difference in the rate of good outcomes (mRS 0–3) between the three treatment regimens (Table 1). Similarly, there was no difference in outcome when we compared subjects that received no corticosteroids or S-DXM with subjects receiving L-DXM (p=0.537). Further, there was no difference in functional disability when we compared subjects who received no corticosteroids versus the combined S-DXM and L-DXM groups (p=0.490). Lastly, on multivariable regression, higher GCS (p<0.001) and lower mFS score (p<0.001), but not DXM treatment regimens (entered separately as S-DXM and L-DXM as well as combined S-DXM/L-DXM), were independently associated with good outcome on discharge (p>0.05, each; Table 2). Additional analysis did not show the effect of S-DMX, L-DXM or S-DXM/L-DXM on the functional disability in groups defined by (1) HH grade on admission (HH 0-II and III-V), (2) presence or absence of IVH,(3) the treatment method and (4) the requirement for EVD placement on admission (p>0.05, each; not shown). After excluding patients who died during hospital stay (n=41) the results did not substantially change (p>0.05, each; not shown).

Long course of DXM was associated with increased rate of infections and hyperglycemia

Overall, patients who received L-DXM had significantly more complications as compared to patients who received no corticosteroids or S-DXM (78.4% vs. 58.6%; p=0.005). There was a difference in the rate of pneumonia, urinary tract infection and hyperglycemia, but not in the rate of ventriculitis/meningitis or delirium between treatment groups (Table 1). After correction for multiple comparisons, patients treated with L-DXM were significantly more often diagnosed with pneumonia as compared to the group treated with no corticosteroids (35.8% vs.13.5%; p=0.010).

On univariate analysis, L-DXM, lower GCS score, and higher mFS score were associated with greater incidence of the primary safety measure as well as infection and hyperglycemia (Table 3). After adjustment, L-DXM remained independently associated with increased risk for the primary safety measure as well as infection and hyperglycemia, respectively (Table 3). When combined, any DXM treatment regimen (S-DXM/L-DXM) was independently associated with an increased risk of infection (OR=3.91; 95%CI, 1.55–9.87; p=0.006), but not hyperglycemia (OR=1.45; 95%CI, 0.66–3.12; p=0.354). There was no significant association between treatment regiments and the risk of delirium in unadjusted and adjusted analyses, respectively (p>0.05 each).

Table 3.

Multivariable logistic regression models for factors associated with the incidence of adverse events attributed to dexamethasone (DXM).

Factors associated with the composite safety outcome (1)
	Crude OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Age (years)	1.02 (0.99–1.04)	0.159	-	-
Female gender	0.85 (0.43–1.67)	0.636	-	-
GCS score	0.76 (0.67–0.86)	<0.001	0.83 (0.73–0.94)	0.003
mFS score	2.19 (1.64–2.94)	<0.001	1.82 (1.31–2.51)	<0.001
*L-DXM	2.56 (1.33–4.92)	0.005	2.72 (1.30–5.72)	0.008
Factors associated with the incidence of infection (2)
	Crude OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Age (years)	1.00 (0.98–1.03)	0.693	-	-
Female gender	1.03 (0.57–1.86)	0.918	-	-
GCS score	0.83 (0.77–0.90)	<0.001	0.86 (0.79–0.94)	0.001
mFS score	1.72 (1.34–2.20)	<0.001	1.41 (1.07–1.86)	0.015
*L-DXM	3.14 (1.59–6.22)	0.001	3.45 (1.63–7.30)	0.001
Factors associated with the incidence of hyperglycemia (3)
	Crude OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Age (years)	1.01 (0.99–1.03)	0.285	-	-
Female gender	1.09 (0.60–1.98)	0.782	-	-
GCS	0.85 (0.79–0.92)	<0.001	0.92 (0.85–1.01)	0.082
mFS	2.11 (1.62–2.73)	<0.001	1.90 (1.43–2.53)	<0.001
*L-DXM	2.06 (1.11–3.81)	0.022	2.05 (1.04–4.04)	0.039
Factors associated with the incidence of delirium (4)
	Crude OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Age (years)	1.02 (0.98–1.05)	0.267	1.02 (0.99–1.06)	0.224
Female gender	0.57 (0.2501.32)	0.189	0.48 (1.20–1.16)	0.103
GCS	1.02 (0.92–1.14)	0.687	1.04 (0.92–1.17)	0.557
mFS	1.03 (0.74–1.45)	0.850	1.08 (0.73–1.60)	0.696
*L-DXM	0.86 (0.35–2.12)	0.751	0.85 (0.34–2.12)	0.849

We used p<0.05 as criteria for backward steps. Blank cells with dashes represent variables that were dropped as non-significant during stepwise selection.

⁽¹⁾Hosmer-Lemeshow goodness of fit χ² 1.43, p=0.964.

⁽²⁾Hosmer-Lemeshow goodness of fit χ² 9.27, p=0.234.

⁽³⁾Hosmer-Lemeshow goodness of fit χ² 7.27, p=0.402.

⁽⁴⁾Hosmer-Lemeshow goodness of fit χ² 10.672, p=0.221. GCS=Glasgow Coma Scale; L-DXM= long course of dexamethasone; mFS=modified Fisher Scale.

^*No steroids and S-DXM (combined) were used as reference group in the model presented in the table.

Discussion

The most important finding of our study was that neither DXM treatment regimen was independently associated with the rate of VPS placement or a good functional outcome on discharge. However, L-DXM was associated with a higher incidence of adverse events, in particular infection and hyperglycemia.

In contrast to majority of previous studies, which explored the effect of corticosteroids on the development of vasospasm and delayed cerebral ischemia^13,19–23, we assessed the effect of DXM on the resolution of hydrocephalus in aSAH. The rationale for using corticosteroids relates to the notion that blood in the CSF elicits a focal inflammatory response leading to impaired absorption in the acute phase of post-SAH hydrocephalus^{24, 25}. Moreover, preclinical studies on the effect of intramuscular methylprednisolone acetate injection showed reduced incidence of hydrocephalus after SAH²⁴. However, we found no difference in the ratio of EVD and VPS placement between different DXM treatment regimens (S-DXM, L-DMX, or any regimen combined) when compared to subjects that did not receive corticosteroids. To date only one clinical retrospective study assessed the effect of corticosteroids on hydrocephalus incidence²⁶. This study suggested that DXM dosed at >12mg/d for at ≥5 days followed by a slow wean may reduce the incidence of hydrocephalus requiring VPS. However, results were confounded by younger age and less severe presentations in the group not requiring VPS placement. Overall, there is presently no compelling evidence to suggest that corticosteroids reduce the risk of hydrocephalus or requirement for short (EVD) or long term (VPS) CSF diversion after aSAH.

We found no effect of DXM on the functional outcome on discharge after adjustment for pertinent confounders. Nevertheless, it is noteworthy that corticosteroids may have a beneficial effect on long-term functional outcomes after aSAH^{13, 22, 26}. Accordingly, our observation period may have been too short to detect a difference between groups. Conversely, the lack of a beneficial effect of corticosteroid therapy in our study may relate to the fact that the majority of patients was treated by endovascular therapy. It has been suggested that the effect of DXM on functional disability after aSAH may differ based on the treatment strategy chosen to secure the aneurysm. For example, high dose DXM (equivalent to L-DXM in our study) was found to be beneficial in micro-surgically treated patients, but not in endovascular group.²² However, in our cohort, surgically treated patients constituted the majority (61.9%) of S-DXM and they were underrepresented (10.1%) in L-DXM group. Arguably, S-DXM was chosen to mitigate surgery-related edema formation rather than for a hypothetical benefit for reducing general inflammatory responses. Yet, our sensitivity analyses, performed separately in surgically and endovascularly treated patients, did not show the effect of DXM on the primary and secondary outcome measures.

An important caveat of corticosteroid therapy is the association with several adverse events including infections, hyperglycemia and delirium, which are independent predictors of a poor outcome after acute brain injury^27–29. Indeed, we found that the use of L-DXM was associated with increased risk of infections and hyperglycemia. Some authors did not find an association of corticosteroid therapy with the rate of infections^{23, 26}; however, the used corticosteroid doses were lower²³ or not well defined²⁶ impairing comparison to our results. In contrast, our findings are consistent with the results of Czorlich et al., who reported a higher incidence of infections with high dose DXM (equivalent to our L-DXM protocol) as compared to no treatment with corticosteroids²². Our S-DXM group was underrepresented (n=21 in S-DXM vs. n=148 in L-DXM) to demonstrate a direct effect of S-DXM on the incidence of infections; however, the rates of pneumonia and urinary tract infection in different regimen groups suggested a duration of DXM treatment dependent effect. Lastly, L-DXM was associated with a higher rate of hyperglycemia in our study, which is consistent with previous reports²². Together with prior observations, our results add to the notion that corticosteroid therapy, particularly in high doses, is associated with an increased risk for adverse events among patients with aSAH.

Strengths of our study relate to the detailed characterization of our cohort with adjustment of our analyses for key clinical and radiological confounders. In addition, we conducted detailed analyses of the regimen dependent effect of DXM on the VPS placement, functional disability at discharge and DXM side effects, which was not previously reported. Lastly, we included patients from two different hospital cohorts which increases the generalizability of our results.

Limitations of our study include those inherent to its retrospective design, for which reason results should be considered hypothesis generating only. Furthermore, all patients who received L-DXM were recruited from one center (UMMMC), therefore the analyses could not be adjusted for the recruitment location due to complete separation of the data. Nevertheless, baseline clinical and radiological characteristics were similar between centers. In addition, due to overall low number of patients treated surgically (n=28), the analysis may have been underpowered to detect the effect of DXM as stratified by the treatment method. Also, due to a low number of patients who met primary outcome measure (VPS placement; n<15), we were not able to assess the effect of DXM on the VPS longevity. Lastly, we collected mRS score at discharge for which reason long-term effects of DXM treatment in aSAH could not be assessed.

Conclusions

The use of DXM among patients with aSAH was not associated with a decrease in the rate of EVD and VPS placement, EVD treatment duration, and early functional outcome improvement, but increased the risk for medical complications. Our observations add to the mounting data, that routine use of DXM in aSAH is not beneficial and may be harmful at high doses.