Health-related quality of life after spontaneous subarachnoid hemorrhage – a prospective cohort study

This observational, prospective cohort study, enrolled patients with a spontaneous SAH who were admitted to the neurological intensive care unit (ICU) of the Medical University of Innsbruck between April 2010 and October 2022. The study period was one year for each patient. Of 498 patients, 183 patients fulfilled the following inclusion criteria: (1) diagnosis of a spontaneous aneurysmal and non-aneurysmal SAH confirmed by computed tomography (CT) scan or lumbar puncture, (2) age greater or equal to 18 years, (3) ICU stay for more than 24 h, (4) fluent in the German language, and (5) multidimensional outcome evaluation including HR-QoL at the 1-year follow-up. The conduct of the study was approved by the hospital’s institutional review board (Medical University of Innsbruck, AM4091-292/4.6). Written informed consent was obtained from all patients according to local regulations, in accordance with the Declaration of Helsinki. Results are reported based on the Strengthening the Reporting of Observational Studies in Epidemiology statement.

All patients were treated according to evidence-based international guidelines, except for nimodipine, which was administered intravenously in poor-grade patients [20‐22]. All patients received prophylactic nimodipine. Ruptured aneurysms were treated by endovascular coiling or neurosurgical clipping based on interdisciplinary case-by-case discussion. Repeated transcranial color-coded duplex sonography (LOGIQ S8; GE Healthcare, Chicago, IL) was performed to detect sonographic large-vessel vasospasm, defined as an increase in mean velocities > 120 cm/s in the middle or anterior cerebral artery or a daily change in mean velocities > 50 cm/s. In cases of severe vasospasm, intra-arterial nimodipine was administered after confirmation by catheter cerebral angiogram. Delayed cerebral ischemia (DCI) was defined as neurological deterioration (new focal neurological deficits and/or a ≥ 2-point decline on the Glasgow Coma Scale) and/or new ischemic lesions on follow-up imaging (CT or magnetic resonance imaging) not attributable to other causes [23].

Patient demographics, disease-specific characteristics, treatment interventions, and hospital complications were prospectively recorded in our institutional SAH database and confirmed at weekly meetings of the study team and attending neurointensivists.

Functional outcome was assessed at discharge, as well as 3 months and 12 months after SAH using the modified Rankin Scale (mRS). The mRS was used as a categorical (mRS ≤ 2 vs mRS > 2) or ordinal level variable and reported accordingly.

One year after SAH, patients underwent a structured interview, clinical examination, neuropsychological testing and completed self-report questionnaires. All of the following outcome instruments were completed one year after SAH:

HR-QoL as assessed by the SF-36 was the primary outcome measure of this study. The SF-36 is a self-report questionnaire that assesses the subjective health condition [11]. It provides scores for eight health domains: Physical functioning, physical role functioning, bodily pain, general health perceptions, vitality, social role functioning, emotional role functioning, and mental health. Each subscale ranges from 0 to 100 points, where higher scores imply better health conditions. The first four subscales are summarized into the physical component score (PCS). The last four subscales are summarized into a mental component summary (MCS). MCS and PCS were rescaled in relation to a 1994 norm-based German population, in which scores were standardised to 50 ± 10 (mean ± SD) [24]. By definition, a normal population has mean scores of 50 with standard deviations of 10. In the current study, norm-based T-scores below 40 were considered to be below the normative range. The Hospital Anxiety and Depression Scale (HADS; self-report questionnaire) was used to detect subjective symptoms of anxiety and depression in SAH survivors [25]. It consists of subscale for anxiety (HADS-A) and for depression (HADS-D), each of which includes seven items scored from 0 to 3. Scores on each subscale range from 0 to 21. Lower scores are associated with less severe anxiety and depression symptoms. Scores > 7 are suggestive of mild to clinically meaningful anxiety or depression disorder and were used for multivariable analysis. Scores > 10 indicate severe anxiety or depression disorder.

Subjectively perceived attention deficits in everyday situations were assessed with the FEDA (Fragebogen Erlebter Defizite der Aufmerksamkeit; self-report questionnaire) [26], which consists of three subscales (27 items): distractibility and retardation in mental tasks (FEDA-1; e.g., watching movies), tiredness and retardation in practical activities (FEDA-2; e.g., doing laundry), and possible reduction in drive (FEDA-3; e.g., interest in hobbies). The FEDA is standardized and scores ≤ 40/65 on the first, ≤ 27/40 on the second, and ≤ 17/30 on the third subscale are below the 10th percentile of the norm and were considered abnormal for the current study [27].

Neuropsychological testing was performed by experienced neuropsychologists, blinded to clinical outcome, one year after SAH using a standardised test battery as described previously. [5]. Executive function was assessed with six different tests: (1) the Frontal Assessment Battery (FAB) [28, 29]; (2) the semantic verbal fluency test (animals per minute)—Regensburger Verbal Fluency Test (Regensburger Wortflüssigkeitstest; RWT) to assess executive language functions [30]; the Trail Making Test (TMT) A (3) and B (4) to assess processing speed and mental flexibility [31]; (5) the digit span backwards subtest of the Wechsler Memory Scale revised assessing verbal working memory [32]; and (6) a clock drawing task (CLOX) to assess figure planning [33]. Executive function impairment was assumed if the patient scored below the 5th percentile (FAB, age-adjusted), the 10th percentile (according to age and/or education stratified norms) or below absolute values (CLOX < 11 points) on 2 out of 6 tests.

Two measures were used to assess memory function: (1) the delayed recall subtest of the Verbal Learning and Memory Test (Verbaler Lern- und Merkfähigkeitstest; VLMT) [34]; and (2) the delayed recall subtest of the Rey–Osterrieth complex figure test (ROCF) [35]. Impairment of memory function was assumed if the patient scored below the 10th percentile (according to age stratified norms) on 1 out of 2 tests.

Attention deficits were assumed if the patient scored below the 10th percentile (according to age norms) on the digit span forwards subtest of the Wechsler Memory Scale revised (WMS-R) [32].

Deficits in visuoconstruction were assessed using the copy subtest of the ROCF [35] and considered to be present if the patient scored below the 10th percentile.

Categorical variables are presented as counts and percentages, and continuous variables are presented as medians and interquartile ranges or means and standard deviations. Based on the distribution of variables, parametric and non-parametric procedures including t-test, Mann–Whitney U test and Kruskal–Wallis test were used. Categorical variables were analyzed using chi-squared and Fisher’s exact test. The Wilcoxon rank test was used to assess differences in functional outcome over the one-year period. Two approaches were used to identify risk factors for reduced HR-QoL: 1) clinically relevant risk factors assessed during the acute phase of the disease were tested, 2) co-occurring multidimensional outcome factors assessed at one year were tested together with acute phase factors. Therefore, clinically meaningful and significantly associated factors (p < 0.1) in the univariate analysis were entered into a multivariable logistic regression model together with predefined covariates (Hunt&Hess, age). Backward elimination was used to remove nonsignificant factors until the best model with the lowest QIC (Quasilikelihood under the Independence model Criterion) values was obtained and the predictors were significant (p < 0.05). Overall reduced HR-QoL, PCS < 40 and MCS < 40 served as binary dependent variables. A p-value less than 0.05 was considered statistically significant. Statistical analysis was performed using IBM SPSS Statistics version 24 64-bit edition.

Of 498 patients admitted to our unit, 89 died in the ICU and a further 22 died during the first year, leaving 387 patients available for potential 1-year follow-up. Functional outcome (mRS) was available for 354 survivors and was comparable to 219 survivors who had detailed follow-up including neuropsychological testing one year after SAH (p = 0.117). Of these 219 patients, 183 completed the HR-QoL survey. The median age was 53 (IQR 46–61) years and 103 (56%) were female. Although there was a shift towards patients with good grades (median Hunt&Hess 2, IQR 2–3; Hunt&Hess 1–3: 84%), patients of all grades were included (Table 1). Included patients were generally younger, less severely affected, had fewer complications and better outcomes compared to 36 patients with detailed one-year outcomes who did not complete the SF-36 (Table 1).

HR-QoL was reduced in 66/183 (36%), while 35 (19%) patients reported restrictions in the physical component summary, 48 (26%) in the mental component summary, and 17 (9%) in both components. Mean scores for both components were below population norms (physical component: 48.38 ± 9.04, p = 0.033; mental component: 46.99 ± 12.31, p < 0.001). Among the subscales, the highest mean difference from population norms was found for physical role (-17.70, p < 0.001), followed by emotional role (-16.38, p < 0.001), social functioning (-6.93, p < 0.001), and physical functioning (-5.50, p = 0.001). Bodily pain, general health, vitality, and mental health were within population norms (Fig. 1, Supplemental Table 1).

Patients with a higher disease severity on admission (higher Hunt&Hess score) were more likely to have reduced physical QoL one year after SAH (PCS < 40, p = 0.040; Supplemental Table 2) but not MCS < 40 (p = 0.572). This association did not persist in multivariable analysis. Supplemental Table 2 shows univariate associations between patient characteristics, hospital complications, short-term outcomes and one-year HR-QoL.

Factors obtained at ICU stay that were associated with reduced overall HR-QoL in multivariable analysis included female sex [adjOR (95%-CI) 2.08 (1.06–4.11), p = 0.034] and a higher discharge mRS [per point, adjOR (95%CI) 1.51 (1.16–1.98), p = 0.029]. A higher mRS at discharge [per point, adjOR (95%CI) 1.86 (1.33–2.60), p < 0.001] was associated with reduced physical QoL (PCS < 40) in multivariable analysis. The only factor associated with reduced mental QoL (MCS < 40) in multivariable analysis was female sex [adjOR (95%CI) 2.22 (1.06–4.62), p = 0.034]. All models were adjusted for age and Hunt&Hess.

Median functional outcomes as assessed by the mRS improved from 2 (IQR 1–3) at discharge to 1 (IQR 0–2) at 3 months (p < 0.001) and further improved to 1 (IQR 0–1) at 1 year after SAH (p < 0.001). At discharge, 57% achieved a good functional outcome (mRS ≤ 2), 87% at 3 months and 96% at one year.

Patients with poor functional outcome at discharge, 3 months and 12 months had lower PCS (p < 0.001, p < 0.001, p = 0.007, resp.) but not MCS scores at one year (p = 0.195, p = 0.520 and p = 0.668, resp., Fig. 2). A discordance between the PCS and MCS of almost 8 and 15 points was found for patients with poor 3 and 12-month functional outcomes, respectively.

Cognitive deficits at the 1-year follow-up, including executive deficits (p = 0.002) and memory deficits (p = 0.003), were more common in patients with reduced HR-QoL (Table 2).

At the one-year follow-up, 31/181 (17%) patients reported symptoms of depression and 61/181 (34%) reported symptoms of anxiety. Severe depression and anxiety (HADS > 10) were evident in 6% and 13% of patients, respectively. Self-reported symptoms of depression and anxiety were more common in patients with reduced HR-QoL including MCS and PCS < 40 (p = 0.001; Table 2).

Increased distractibility in everyday situations (FEDA-1 subtest) was reported by 20% of patients, increased tiredness (FEDA-2) by 39%, and a reduction of drive (FEDA-3) by 13%. All three symptoms were more common in patients with reduced HR-QoL including MCS and PCS < 40 (p = 0.001; Table 2).

In multivariable analysis, including variables from the acute phase of the disease, 3-month and 12-month multidimensional outcome variables, a lower Hunt&Hess score at admission (p = 0.036), female sex (p = 0.017), self-reported depression (HADS-D > 7, p = 0.001), increased tiredness (FEDA-2, p < 0.001), and a reduction of drive (FEDA-3, p = 0.019) one year after SAH were associated with overall reduced HR-QoL and explained 68.9% of the variance (Table 3, Fig. 3). Self-reported anxiety and depression (HADS-A > 7, p < 0.001; HADS-D > 7, p = 0.029), increased tiredness (FEDA-2, p = 0.009), and a reduction of drive (FEDA-3 p = 0.003) at one year were associated with reduced MCS. These factors explained 61.7% of the variance in reduced MCS. Worse three-month functional outcome (higher mRS, p = 0.001) and increased tiredness at one year (FEDA-2, p < 0.001) were associated with reduced PCS, explaining 53.9% of the variance (Fig. 3). All models were adjusted for age and Hunt&Hess score on admission.