The cohort has been reported previously (16, 22, 23). A total of 26 consecutive patients, treated for acute non-traumatic, angiographically confirmed aneurysmal or non-aneurysmal perimesencephalic SAH (pSAH) at our University Medical Center between February 2013 and May 2016, fulfilled the beforehand specified eligibility criteria of being native German speakers, aged 18–75 years, who had been admitted to hospital within 48 h of ictus in prognostically favorable, good to moderate neurological condition, defined as a World Federation of Neurological Surgeons (WFNS) score (24) and Hunt and Hess (HH) grading (25) of 1–4, respectively, and an initial Glasgow Coma Scale (GCS) of ≥9. Written informed consent was obtained from each individual. In all patients, the diagnosis of sSAH was based on cerebral computed tomography (CT) and a four-vessel digital subtraction angiography (DSA) to visualize the location and morphology of the underlying ruptured aneurysm and to differentiate aneurysmal from perimesencephalic hemorrhages, respectively. Within the first 72 h after the onset of sSAH, all patients received an external ventricular drain (EVD) due to a radiologically confirmed acute occlusive hydrocephalus. Treatment consisted either of endovascular aneurysm occlusion (EV group) or microsurgical clipping (MS group) for aSAH patients, followed by standardized therapy in the intensive care unit (ICU) (pSAH group). Our standardized microsurgical and endovascular procedure protocols (26) and our ICU standard operating protocol (27) have been described elsewhere.

The clinical database comprised all demographic, radiological, and neurological variables, comorbidities, non-/invasive procedures, complications, comprehensive pharmacological screening (at discharge and at the 6-month FU), and outcome grading [Glasgow Outcome Scale (GOS) (28) and modified Ranking Scale (mRS) (29)].

Subacutely (between day 11 and 35; t₁) and 6 months (t₂) after the onset of sSAH, all patients completed two well-established generic self-report measures of physical and mental health each: The ICD-10-Symptom-Rating questionnaire (ISR) (30) serves as a score for symptom burden and the German version of the 36-Item Short Form Health Survey (SF-36) (31) represents a performance score.

Within the 10-day period after the onset of sSAH, plasma was drawn directly from the arterial line once daily. At 3 weeks and at the 6 month-FU, blood was collected by venipuncture. Immediately subsequent to sampling, the samples were centrifuged at 1,200 rounds per minute for 10 min, and the supernatants were aliquoted and stored at−80°C until further use. The samples were thawed, aliquoted (1 ml) with columns [STRATA C18-E (55 μm, 70 A) 200 mg/3 ml-columns, Phoenix Pharmaceuticals Inc., Burlingame, USA], purified, evaporated on a vacuum concentrator (Christ RVC 2-25 CD plus; Osterode am Harz, Germany), and dissolved in 250 μl assay-buffer resulting in a fourfold concentration. CGRP levels were measured in duplicate purified serum samples using a competitive enzyme immunoassay (EIA; Phoenix Pharmaceuticals Inc., Burlingame, CA, USA). The cerebral exposure to the endogenously released CGRP into serum over time is expressed as ng/ml.

Continuous data and neuropsychological test results are presented as mean ± standard deviation (SD) and range (minimum to maximum) and categorical data as frequency counts.

Neuropsychological assessment: Changes over time within each group were analyzed with a paired t-test. Differences between groups at postinterventional assessment were analyzed with an analysis of variance (ANOVA) followed by Fisher's LSD post-hoc pairwise comparisons.

Correlation of CGRP with neuropsychological assessment: Regression analyses were conducted for correlations of mean CGRP levels with the hrQoL test scores. Changes over time within each group were analyzed with a paired t-test. Intragroup variances (correlations between hrQoL test scores and clinical variables) were analyzed using an analysis of variance (Bartlett's test for equal variances). Statistical analysis was conducted according to Stata procedures (Stata Version 14.2; Stata Corp. College Station, TX, USA).

A p < 0.05 was considered statistically significant. A biostatistician was involved in the design and realization of the statistical evaluation.

A total of N = 160 patients with sSAH were treated during the observation period. In accordance with our strict selection criteria, depicted in the Flowchart (Figure 1), five patients were excluded from analysis. Thus, a total of 21 patients (8 men, 13 women) of a mean age of 50.6 years (range 27–72 years) with good-grade sSAH was assigned to statistical evaluation. Four aSAH patients, initially presenting with HH grade III (14%) or IV (5%), neurologically improved immediately after insertion of an EVD following hospital admission (i.e., HH I or HH II). Thus, the poorer HH score was obviously related to acute occlusive hydrocephalus. With this qualification, we consider the term “good-grade” sSAH patients as appropriate for our cohort. An aneurysmal bleeding source in the anterior (n = 12) or posterior circulation was detected in 15 patients (EV n = 9: coil n = 5, stent-assisted coil n = 2; balloon-assisted coil n = 1; flow diverter n = 1). No patient developed serious procedure-related complications like, for example, periprocedural aneurysm rupture, procedure-related blood transfusion, postprocedural onset of a new neurological deficit, or clinically symptomatic CV-related stroke. Functional outcome was stable or even improved over the 6-month FU. None of these patients required decompressive craniectomy, and, until FU, no late rebleedings or mortalities had occurred. Each individual was able to complete both of our surveys, the ISR and the SF-36, within 10–15 min, and without any signs of (mental or physical) fatigue during the testing, neither at t₁ nor at t₂. Comprehensive information on the baseline data including the aneurysm site, GCS, HH, WFNS and Fisher score, procedure variables, medication, and outcome grading is provided by Table 1.

The presented sSAH cohort exclusively encompassed patients with severe radiological grades of sSAH (Fisher score 3 or 4) and, consecutively, is imbalanced in terms of the overestimation of the true hydrocephalus rate after sSAH. This is all the more remarkable given the inclusion of 6 patients with pSAH. Only four patients, however, required permanent ventriculoperitoneal shunting. Statistical intergroup comparisons yielded no significant differences except for a higher number of middle cerebral artery aneurysms in the MS group (p = 0.022), a higher intake of antiplatelets in the EV group (p = 0.016) at t₁, an unsurprisingly longer duration of MS vs. EV (p = 0.004), and a longer mean time spent on mechanical ventilatory support in the EV group than in the pSAH group (EV vs. pSAH p = 0.0496; EV vs. MS p = 0.864, MS vs. pSAH p = 0.065).

The self-reported hrQoL perception has been reported previously (22, 23) sum up, our cohort performed significantly worse in several ISR- and SF-36-subscales than the particular population norms, especially at t₁. Within 6 months, sSAH patients had significantly improved with regard to bodily pain (Pain), physical functioning (Pfi), the physical component summary (PCS), depression, anxiety, general health perceptions (Ghp), and social functioning (Social). Poor self-reported neuropsychological performance [physical SF-36 items: Pain, role limitations because of physical health problems (Rolph), Ghp; ISR scores: total, depression, compulsive-obsessive syndrome] in the subacute phase correlated with a worse outcome on the GOS at discharge. And the HH score correlated positively with all psychological SF-36 item scores [vitality (Vital), Social, role limitations because of emotional problems (Rolem), general mental health (Mhi), mental component summary (MCS)] at t₁.

During the first 10 days, the mean CGRP levels in serum (0.470 ± 0.10 ng/ml) were significantly lower than the previously (16) analyzed mean CGRP values in CSF (0.662 ± 0.173; p = 0.0001). The mean serum CGRP levels within the first 10 days did not differ significantly from the values at 3 weeks (0.493 ± 0.163 ng/ml; p = 0.304). At 6 months, the mean serum CGRP value (0.429 ± 0.121 ng/ml) was significantly lower compared to 3 weeks (p = 0.010) and compared to the first 10 days (p = 0.026) (cf. Figure 2).

In female patients, CGRP levels at the 6-month FU were significantly higher than in male patients (0.473 ± 0.121 ng/ml vs. 0.358 ±0.087 ng/ml; p = 0.016). Further regression analyses did not reveal any significant correlation between the neuropeptide levels and other patient variables like, for example, the treatment modality.

Higher mean serum CGRP levels at 3 weeks (p = 0.001) and at the 6-month FU (p = 0.005) correlated with a significantly poorer performance in the SF-36 item pain, and, at 3 weeks, with a higher ISR symptom burden regarding somatoform syndrome (p = 0.001) at t₂ (cf. Figure 3). The analysis of cognitive test performances is summarized in Table 2.

Compared to other life-threatening medical events, stroke has been proven to be associated with the strongest reductions in hrQoL and functioning status (32). In stroke patients, the physical hrQoL increased after treatment (except for bodily pain), whereas the psychological quality remained low (33). This equals our results in sSAH which indicate a significant improvement in multiple physical items (Pfi, Pain, Ghp, PCS) within 6 months after sSAH, while most psychological items (Compulsive-obsessive syndrome, somatoform syndrome, nutrition disorder, the ISR supplementary items, the ISR total score, Vital, Rolem, Mhi, and MCS) remained stable.

The trigeminovascular system is involved in the regulation of the cranial vasculature and represents a key element in the transmission of pain (47). Upon activation in primary headaches, in stroke (48), and in SAH (15, 16, 44, 49), the trigeminovascular system, a vasodilator pathway, antidromically releases the peptide neurotransmitter CGRP (13), hereby increasing the cerebral blood flow (CBF) via vasodilation (13) and mediating pain by stimulation of the trigeminus nucleus caudalis, respectively (47). The activation of this (putative defense) system is noted clinically as a marked increase in the cranial venous outflow of CGRP during headache attacks or in SAH, where it was elevated in the CSF as well (14). The psychoactive mediator (17) CGRP has been repeatedly attested a crucial involvement in a variety of neurobehavioral and psycho-affective conditions (18), like in depression, anxiety and learning and memory [cf. (16)], as well as in inflammatory and neuropathic pain (19, 50), and, by pronounced arterial cerebral vasodilation, in migraine (20). Congruently, in our series, elevated CGRP levels in serum at 3 weeks and at 6 months significantly correlated with more pain and, at 3 weeks, with a higher symptom burden regarding somatoform syndrome 6 months after sSAH. We qualify the statement with the note that our utilized set of standardized and validated physical and mental health questionnaires refers to bodily pain in general and is therefore incapable to establish a headache-specific diagnosis. However, in the semi-structured interview during FU assessment recurrent headache was a commonly reported symptom in our series. To what extent CGRP is involved in non-headache conditions has not been clarified yet. A recent review article (19) revealed an association between measured CGRP levels and somatic visceral, inflammatory and neuropathic pain. Upregulated CGRP levels were reported in serum, CSF, synovial and tissue biopsies in patients with degenerative disc disease, osteoarthritis, and temporomandibular joint pain. Furthermore, CGRP was elevated in acute pain conditions and pain after exercise. In somatic pain conditions in particular, CGRP levels correlated with pain. This is in concert with our results as to somatoform syndrome after sSAH. Growing evidence indicates that CGRP facilitates nociceptive transmission and contributes to the development and maintenance of a sensitized, hyperresponsive state of the primary and second-order afferent sensory neurons, thus contributing to peripheral and central sensitization. The maintenance of such a sensitized neuronal condition is believed to be a key player underlying migraine pathophysiology (50). In sSAH, the (potentially adverse) effects of CGRP on pain modulation and transmission have not been explored before. Our data conflicts with the substantiated beneficial cerebroprotective role of (endo- and exogenously administered) CGRP in hemodynamics (14, 15, 46).

Upon nerve stimulation, CGRP is released from its storage vesicles in sympathetic perivascular fibers, which form a particularly dense network around the major arteries of the anterior part of the circle of Willis (14), and in free nerve endings in the dura mater (51). Pathophysiologically, it is supposed that, at the moment of aneurysm rupture, these CGRP-containing nerve fibers are affected directly or indirectly–by the stimulation of the vessel's wall or surrounding nociceptors or cortical neuronal depolarization secondary to the hemorrhage or by the blood in the subarachnoid space–initiating activity of the trigeminovascular system and inducing an excessive release of CGRP from the perivascular nerve terminals into CSF and serum (14, 15, 44). Post mortem analyses after SAH and early observational studies on the cerebral circulation after experimental SAH confirmed a marked decrease in CGRP immunoreactivity in the perivascular nerve fibers, accordingly [cf. (46)]. To date, our study is the first to provide an insight into the temporal dynamic changes of CGRP in serum after good-grade sSAH. Within the first 10 days and 3 weeks after the onset of hemorrhage, CGRP levels did not differ significantly. In contrast, 6 months thereafter, serum CGRP values had significantly declined. Data on the time course of CGRP in the serum is scarce. The most marked suppression of CGRP immunoreactivity in cerebrovascular nerve fibers was documented during the 7th to 14th day after artificial SAH in dogs with a recovery to normal levels by the 42nd day (52). Hypothetically, the CGRP release is followed by an inhibition of the CGRP reuptake at the nerve-ending terminals. CGRP-receptor complexes are formed, and, possibly after a non-competitive saturation of the extra- and intraluminal CGRP receptors, CGRP levels in CSF and serum finally decrease due to the subsequent depletion of the releasing terminal nerve endings. It has further been speculated that different types of stroke like sSAH may alter the synthesis, metabolism, molecule, or receptors of the neurotransmitter CGRP. In the serum of sSAH patients, peak concentrations of CGRP have been measured after rupture of aneuryms of the middle cerebral artery (14, 44) (during days 4–7 after onset of sSAH), in patients with CV-related ischemia (on day 4), and – as to cerebrovascular manipulation–after endoluminal aneurysm treatment via coiling (44) (within days 3–5 after sSAH). In our series, the serum CGRP levels did neither differ regarding the anatomical localization of the ruptured aneurysm, nor regarding the treatment-induced mechanical intra- and extraluminal manipulation of the parent vessel in the aneurysm-securing procedure groups and the control group, respectively. Interestingly, in our cohort without gender bias, the female patients expressed significantly higher serum CGRP levels at the 6-month FU than their male counterparts. Although headaches are prevalent in both sexes and in all age groups, women aged 20 up to 50 years are those who have the highest prevalences in Europe (53).

A remarkable and presumably cerebroprotective hypersecretion of CGRP into serum counterbalancing vasoconstriction was previously detected in sSAH patients with vasospasm-related ischemia (44). In our cohort, exclusively comprising good-grade sSAH patients without any CV-related ischemic complications, the excessive release of CGRP into CSF (15, 16) during the first 10 days after ictus contrasts with the significantly lower CGRP levels in serum. This is all the more noteworthy given that even these smaller amounts of CGRP in serum suffice to induce considerable impairment in pain processing and psychosomatic disorders. The sparse neuropeptide findings in serum presumably result from the diluting effect of the nanomolar concentrations of the ventricular CSF CGRP during its absorption into plasma and its clearance over time. On the other hand, our results in a good-grade sSAH collective contradict previous findings in aSAH patients with CV (14, 44), in whom serum CGRP levels have been measured higher than CGRP levels in CSF. To date, it is unclear which source is the best to (non-invasively) access predictive CGRP values, like, recently, in tear fluid (53) of migraine patients.

After three decades of intense research, drugs targeting the trigeminal sensory neuropeptide CGRP or its receptor and, thus, acting on the trigeminal pain system have been specifically designed (21). It is remarkable that, in migraineurs, all CGRP-targeted therapies (CGRP receptor antagonists/gepants and monoclonal antibodies against CGRP or the CGRP receptor) have proven clinical efficacy, whether in preventing attack onset, targeting acute attacks, or reducing attacks in chronic migraine or frequent episodic migraine (21, 54). To date, it is unknown whether CGRP antagonism may provide a beneficial therapeutic approach for the treatment of chronic pain conditions and CV in sSAH as well.

Scrutinizing the frequency and type of self-reported met and unmet needs of SAH survivors 1–2 years and 3–5 years following hemorrhage, headache, fatigue, concentration, memory, and anxiety were the most commonly described, and a large proportion of needs (> 80%) were detected as unmet (55). The pain conditions, real-world deficits, and the neurobehavioral sequelae of an sSAH are not appropriately addressed by current rehabilitation programs, which still mainly focus on physical rehabilitation. Neuroscientists and clinicians should strive for a standardized screening and early identification of patients at-risk and for non-pharmacologic coping strategies that may be effective in pain management and hrQoL improvement to increase patient resiliency (56). Comprehensive rehabilitation programs imply a multidisciplinary approach (57) with physiotherapy, cognitive behavioral therapy, and biofeedback using relaxation techniques in order to reduce pain conditions; in addition, it is recommended to screen for and treat analgesics overuse (38).

In methodological terms, the main strength of our investigation is the prospective, controlled data collection comprising a throughly selected cohort of a representative neurovascular medium volume center. The study population underwent a meticulous liquid biobanking in combination with a standardized comprehensive hrQoL assessment over a considerable FU period of 6 months. Attributed to the abrupt nature of sSAH, biomarkers are not amenable to baseline data acquisition. Furthermore, the utilized hrQoL questionnaires are not designed to precisely establish pain and headache diagnoses and to specify the psychosomatic symptom burden. A major limitation of our experimental clinical study is the circumscribed sample size demanding a cautious interpretation of our findings. We are aware of the potential bias due to the exclusion of five patients. The results are not applicable for sSAH patients in general since our cohort exclusively encompasses prognostically favorable good-grade individuals with the per-protocol exclusion of non-HC patients.