Multimodal neuromonitoring-related complications in patients with severe acute brain injury: a retrospective cohort study Authors: Tobias Sjørslev Bodilsen, Trine Hjorslev Andreasen, Markus Harboe Olsen, Anne-Sophie Worm Fenger, Alexandra Vassilieva, Helene Ravnholt Jensen, Kirsten Møller Published: 2026 ABSTRACT BackgroundInvasive multimodal neuromonitoring (MMNM) is standard of care in many neuro-intensive care units for patients with severe acute brain injury. MMNM is used for monitoring of early signs of secondary brain injury. Still, the safety of this approach is only partly clarified. This study aimed to investigate the prevalence and types of MMNM-related complications.MethodsFor this single-centre retrospective study, we included all patients admitted to the Neurointensive Care Unit at Rigshospitalet from March 2017 to December 2023, who were monitored using MMNM probes in the brain, i.e. monitoring of intracranial pressure (ICP) in combination with brain tissue oxygen tension (PbtO2) and/or cerebral microdialysis (CMD). Medical records and computed tomography (CT) or magnetic resonance imaging (MRI) of the head were reviewed for MMNM-related complications and functional outcome at 6 months after ictus.ResultsA total of 223 patients with severe acute brain injury underwent monitoring with MMNM. Imaging was done after insertion in 196 patients (88%) of all who received MMNM. In these, haemorrhage related to the insertion was observed in 25 patients (13%), suboptimal location in 70 (36%), bone fragments in 25 (13%), and pneumocephalus in 71 (36%); none of these necessitated intervention. Among the entire cohort of 223 patients, infection related to MMNM occurred in one patient (0.4%); malfunction was reported by clinicians in 64 patients (29%) and verified in 26 (12%). After MMNM removal, 162 patients (73%) had a CT or MRI scan, of whom 11 patients (6.8%) had a new parenchymal haemorrhage; one patient (0.4%) underwent surgical haematoma evacuation. Six months functional outcome did not differ between patients with and those without MMNM-related complications.ConclusionIn this retrospective study, a high number of MMNM-related complications was found, of which the vast majority did not cause clinical concern and did not affect patient treatment or outcomes.Supplementary InformationThe online version contains supplementary material available at 10.1007/s00701-026-06950-z. Introduction Severe acute brain injury, such as traumatic brain injury (TBI), aneurysmal subarachnoid haemorrhage (SAH) and spontaneous intracerebral haemorrhage (ICH), is associated with high morbidity and mortality [7, 13, 24]. These patients usually require admission to an intensive care unit (ICU) with sedation and mechanical ventilation in the acute phase. However, sedation may hinder clinical assessment and detection of neurological worsening. Various combinations of multimodal neuromonitoring (MMNM) are used in Neuro-ICUs worldwide to detect and treat early signs of secondary brain injury, which is associated with unfavourable functional outcome [12, 17, 22]. Some modalities are invasive, requiring a burr hole in the skull to insert catheters for, e.g., monitoring of intracranial pressure (ICP), brain tissue oxygen tension (PbtO2) as well as cerebral microdialysis (CMD) [9]. The procedure incurs a risk of haemorrhage, infection, and suboptimally located catheters, the prevalence and consequences of which are only partially clarified [1, 3, 5, 10, 21]. The aim of this retrospective, single-centre, cohort study of patients with severe acute brain injury was to (1) investigate the prevalence and type of complications due to MMNM, (2) explore potential clinical risk factors of different types of complications, and (3) evaluate the relationship between MMNM-related complications and functional outcome. In accordance with the clinical practice in our institution, MMNM was defined as simultaneous use of at least two of the following modalities: ICP monitoring, PbtO2 monitoring, and CMD. Methods In this single-centre, retrospective cohort study, we included all patients with severe acute brain injury who were admitted to the Neuro-ICU at Copenhagen University Hospital–Rigshospitalet, Denmark from March 2017 to December 2023 and monitored with at least two of the following modalities: ICP, PtbO2, and CMD. Complications related to external ventricular drain (EVD) was not included. The study was approved by the Centre of Regional Development, the Capital Region of Denmark (ref. R-23052162, 20 September 2023), who waived the need for written consent from patients for this type of study according to the Danish Health Act §46 and §47. Indications, MMNM placement and devices During the study period, the local guidelines recommended MMNM with ICP, PbtO2 and CMD for patients admitted with TBI or SAH and a Glasgow Coma Scale (GCS) score < 9. Patients with ICH who were included in the ongoing KETA-BID trial [2] were also eligible to receive MMNM for research purposes. The local guidelines further recommended that MMNM be placed contralaterally to the most injured hemisphere as detected on the last CT scan before surgery in patients with TBI, and ipsilaterally to the ruptured aneurysm or, in case of a midline aneurysm, in the left hemisphere in patients with SAH. In ICH patients, the MMNM was intended to be placed ipsilaterally to the haemorrhage. However, the insertion of MMNM as well as the side of insertion ultimately depended on the neurosurgeon’s discretion and technical feasibility. Insertion of MMNM was done either in the operating room or bedside in the Neuro-ICU. The surgical procedure was performed under general combined with local anaesthesia, using a point incision followed by creation of a burr hole with a manual drill. Next, the dura was penetrated, and a bolt was inserted into the skull. Finally, the catheters were inserted intraparenchymally through the bolt. In the period from March 2017 to June 2023, ICP monitoring was done using an intraparenchymal transducer, Codman Microsensor® ICP Transducer (Integra® Lifesciences, New Jersey, United States) connected to Codman® ICP Express™ monitor (Integra® Lifesciences, New Jersey, United States). PbtO2 monitoring was done using a Licox® Brain Tissue Oxygen Monitor Complete Probe Kit (Integra LifeSciences, New Jersey, United States) connected to Licox® PbtO2 Monitor (Integra LifeSciences, New Jersey, United States). These devices were linked through a double bolt (IM2SEU) or triple bolt (IM3SEU). From June 2023 to December 2023 a Bolt kit PTO 2L (Raumedic, Helmbrechts, Germany) was used with a Neurovent-PTO 2L Bolt (Raumedic, Helmbrechts, Germany) for ICP, PbtO2, Brain tissue temperature. Throughout the period, CMD monitoring was carried out using a 70 MD Bolt Catheter (length, 130 mm, diameter, 0.9 mm; M Dialysis AB, Stockholm, Sweden). Data collection and definitions Demographic data and information on diagnosis, duration of MMNM, and Neuro-ICU and hospital lengths of stay were collected from medical records. Two authors (TSB and THA) independently assessed computed tomography (CT) scans and magnetic resonance imaging (MRI) of the head for complications as well as using the neuroradiologists’ reports. Any disagreement was resolved by consensus. All CT and MRI scans were performed only when clinically indicated. A priori, we defined the following as surgical complications: intracranial haemorrhage, bone fragments, and pneumocephalus regardless of size, and suboptimal catheter location after insertion. Excluding catheter location, surgical complications were further categorised as occurring after catheter insertion or after catheter removal. Surgical complications after removal were only noted if they were new compared to after catheter insertion. The following other events related to MMNM were collected from the medical records: dislodgement, infection that could be related to an MMNM device, leakage and catheter malfunction. Surgical complications after insertion and any other events are collectively referred to as MMNM-related complications. Additionally, MMNM-related complications requiring interventions were noted. Intracranial haemorrhage Intracranial haemorrhage was subcategorised as subdural (SDH) or intracerebral (ICH). The maximal width of SDH was measured. ICH after catheter insertion was further classified as Grade I (< 1 ml), Grade II (1 ml) or Grade III (> 1 ml) for haematoma located along or at the tip of the catheter [1]. ICH volume was calculated using the ABC/2 formula, based on maximal measurements in centimetres in the A, B, and C planes obtained from multiplanar reformations [4]. Infection Skin infection was defined as clinical signs of infection with or without a positive bacterial culture more than 48 h after MMNM insertion, in combination with administration of antibiotics on this indication. Intracranial infection was diagnosed from the medical records if intrathecal and/or systemic antibiotics were prescribed, and suspected or verified intrathecal infection due to MMNM was noted as the indication. Catheters were only sent for culture when clinically indicated. Catheter location Optimal catheter location was defined if the tip of the catheter was visible on CT in normally appearing subcortical parenchyma, as intended, in the frontal lobe without touching other intracerebral devices. Suboptimal catheter location was therefore defined as a catheter located elsewhere, further subcategorised as extraaxial, in the grey matter, within a lesion, perilesional, intraventricular, paraventricular, or in the midline, or if the catheter was kinked or the tip appeared to be in physical contact with e.g. an extra ventricular drain. Catheter malfunction A malfunction was considered to be verified if it was confirmed by a clinician trained to interpret data from MMNM with or without functional validation. Functional validation of PbtO2 catheter was done by increasing the fraction of inspired oxygen (FiO2) to 100% for minimum of 5 min. If PbtO2 did not rise substantially within 10 min during the oxygen increase, despite an increase in peripheral oxygen saturation, the probe was deemed to be non-functioning. CMD malfunction was defined by absence of fluid in the microdialysis chamber. A non-verified malfunction was defined as a report of catheter malfunction entered in the medical record that was not validated by a clinician who had received specific training in MMNM. Interventions Interventions were defined as reinsertion of catheter and/or bolt, including reinsertion through a new burr hole, and surgery or removal of MMNM equipment due to any MMNM-related complication. Functional outcome Functional outcome at 6 months was recorded using the Glasgow Outcome Scale-Extended (GOSE) and modified Rankin Scale (mRS) by two researchers independently and collected by reviewing medical records. An unfavourable functional outcome was defined as GOSE 1–5 and mRS 3–6. All data were entered and stored in REDCap (Version LTS 13.7.14 Vanderbilt University, Tennessee, United States of America). Statistical analyses Analyses were performed using the statistical software R (R version 4.5.0, R Core Team 2024, Vienna, Austria). Continuous variables were assessed for normal distribution with the Shapiro–Wilk test and visually with QQ plots. If normally distributed, data was presented as the mean and standard deviation (SD); otherwise, the median and interquartile range (IQR) were reported. Categorical variables were reported as numbers (n) and percentages (%). The Wilcoxon test and Fisher’s exact test were performed to compare complications between groups. Logistic regression was performed to investigate odds ratio (OR) for selected complications. A p-value < 0.05 was considered statistically significant. Results Patient characteristics and MMNM-related procedures Over an almost 7-year period, 223 patients were monitored with MMNM, of whom 6 were under 18 years of age (Table 1; for comorbidities and Neuro-ICU infection, please refer to Supplementary Table 1 and for severity scoring of patients with SAH, please refer to Supplementary Table 2). Median time from admission to MMNM insertion was 0 days (IQR: 0–1). Insertion of MMNM was performed in the operating room in 188 patients (84%), and bedside in the Neuro-ICU in 35 patients (16%). Seventy-six of 117 patients (65%) with TBI underwent craniotomy or craniectomy with haematoma evacuation. In SAH patients, craniotomy was performed in 35 of 102 cases (34%), one of which was combined with endovascular treatment; 54 patients (53%) were treated endovascularly only, and treatment data were missing for 13 patients. All 4 patients with ICH underwent a craniotomy with haematoma evacuation. A total of 196 patients (88%) had a CT scan performed (median of 2 days, IQR: 1–4) after insertion. Reinsertion of one or more catheters was done for 37 patients (17%), of whom six needed a second reinsertion; a total of 43 reinsertions were performed of which 25 (58%) were followed by a CT scan. MMNM equipment was removed in part or in full and without reinsertion in 43 patients (19%) due to an MMNM-related complication. MMNM modalities and the anatomical location of MMNM devices are summarized in Table 2 and 3. Table 1Patient characteristicsOverallNo MMNM complicationsMMNM complicationsP-value*(N = 223)(N = 58)(N = 165)Demographics Female109 (49%)34 (59%)75 (46%)0.09 Age (years)55 [40–65]57 [40–64]53 [39–65]0.70Diagnosis ICH4 (1.8%)1 (1.7%)3 (1.8%)0.20 SAH102 (46%)32 (55%)70 (42%) TBI117 (53%)25 (43%)92 (56%)Comorbidities$74 (33%)21 (36%)53 (32%)0.63Admission Initial GCS6 [3–12]6 [3–12]6 [3–12]0.69 In-hospital GCS3 [3–10]3 [3–12]3 [3–9]0.29 Charlson score1 [0–3]1 [0–3]1 [0–3]0.98 APACHE II score21 [18–25]21 [18–25]22 [18–25]0.69External ventricular drainage102 (46%)21 (36%)81 (49%)0.09 Combi-drain99 (44%)30 (52%)69 (42%)0.22Duration of MMNM (days)10 [6–16]9 [3–16]11 [6–17]0.05Length of ICU stay (days)16 [6–29]14 [5–28]16 [8–29]0.24Length of hospital stay (days)46 [15–88]48 [18–90]46 [15–88]0.95Infection$161 (72%)39 (67%)122 (74%)0.39Glasgow Outcome Scale-Extended Unfavourable outcome (1–5)166 (74%)41 (71%)125 (76%)0.60 Favourable outcome (6–8)48 (21%)14 (24%)34 (20%) Missing9 (4.0%)3 (5.2%)6 (3.6%)Modified Rankin Scale Unfavourable outcome (3–6)152 (68%)39 (67%)113 (68%)0.88 Favourable outcome (0–2)62 (28%)16 (28%)46 (28%) Missing9 (4.0%)3 (5.2%)6 (3.6%)6 months mortality Dead68 (31%)15 (26%)53 (32%)0.48 Alive150 (67%)41 (71%)109 (66%) Missing5 (2.2%)2 (3.4%)3 (1.8%)Numbers are count (percentage) or median [IQR]APACHE acute physiology and chronic health evaluation; Combi-drain, external ventricular drainage draining cerebrospinal fluid and measuring intracranial pressure simultaneously, GCS Glasgow Coma Scale, ICH Intracerebral haemorrhage, ICU intensive care unit, MMNM multimodal neuromonitoring, SAH subarachnoid haemorrhage, TBI traumatic brain injury*Fisher’s test for categorical variables, Wilcoxon’s test for continuous variables$ Please see supplement 1 for more detailsTable 2MMNM modalities used (N = 223)ModalitiesN (%)PbtO2 and ICP*6 (2.7%)PbtO2 and ICP33 (15%)PbtO2 and CMD96 (43%)PbtO2, ICP and CMD71 (32%)PbtO2, ICP, Tbt and CMD17 (7.6%)CMD cerebral microdialysis, ICP intracranial pressure, PbtO2 brain tissue oxygen tension, Tbt Brain temperature*In 6 patients, no ICP catheter was inserted through the double bolt (IM2SEU) due to surgical difficulties, and contrary to the intention. Thus, ICP was measured using an EVD combi-drain or a tunnelled ICP catheter, and only the PbtO2Table 3Anatomical location of MMNM equipment by clinical diagnosis (N = 223)*ConditionHemisphere locationN (%) of patientsSAH (non-midline aneurysm)Contralateral hemisphere44 (20%)Ipsilateral hemisphere44 (20%)SAH (midline aneurysm)Right hemisphere8 (3.5%)Left hemisphere3 (1.3%)Aneurysm not foundRight hemisphere2 (0.9%)Left hemisphere1 (0.4%)TBIMost injured hemisphere41 (18%)Contralateral hemisphere76 (34%)ICHContralateral hemisphere2 (0.9%)Ipsilateral hemisphere2 (0.9%)*At first placement. Subsequent insertion generally on same side, except for three patients with TBIICH intracerebral haemorrhage, SAH subarachnoid haemorrhage, TBI traumatic brain injury Table 1Patient characteristicsOverallNo MMNM complicationsMMNM complicationsP-value*(N = 223)(N = 58)(N = 165)Demographics Female109 (49%)34 (59%)75 (46%)0.09 Age (years)55 [40–65]57 [40–64]53 [39–65]0.70Diagnosis ICH4 (1.8%)1 (1.7%)3 (1.8%)0.20 SAH102 (46%)32 (55%)70 (42%) TBI117 (53%)25 (43%)92 (56%)Comorbidities$74 (33%)21 (36%)53 (32%)0.63Admission Initial GCS6 [3–12]6 [3–12]6 [3–12]0.69 In-hospital GCS3 [3–10]3 [3–12]3 [3–9]0.29 Charlson score1 [0–3]1 [0–3]1 [0–3]0.98 APACHE II score21 [18–25]21 [18–25]22 [18–25]0.69External ventricular drainage102 (46%)21 (36%)81 (49%)0.09 Combi-drain99 (44%)30 (52%)69 (42%)0.22Duration of MMNM (days)10 [6–16]9 [3–16]11 [6–17]0.05Length of ICU stay (days)16 [6–29]14 [5–28]16 [8–29]0.24Length of hospital stay (days)46 [15–88]48 [18–90]46 [15–88]0.95Infection$161 (72%)39 (67%)122 (74%)0.39Glasgow Outcome Scale-Extended Unfavourable outcome (1–5)166 (74%)41 (71%)125 (76%)0.60 Favourable outcome (6–8)48 (21%)14 (24%)34 (20%) Missing9 (4.0%)3 (5.2%)6 (3.6%)Modified Rankin Scale Unfavourable outcome (3–6)152 (68%)39 (67%)113 (68%)0.88 Favourable outcome (0–2)62 (28%)16 (28%)46 (28%) Missing9 (4.0%)3 (5.2%)6 (3.6%)6 months mortality Dead68 (31%)15 (26%)53 (32%)0.48 Alive150 (67%)41 (71%)109 (66%) Missing5 (2.2%)2 (3.4%)3 (1.8%)Numbers are count (percentage) or median [IQR]APACHE acute physiology and chronic health evaluation; Combi-drain, external ventricular drainage draining cerebrospinal fluid and measuring intracranial pressure simultaneously, GCS Glasgow Coma Scale, ICH Intracerebral haemorrhage, ICU intensive care unit, MMNM multimodal neuromonitoring, SAH subarachnoid haemorrhage, TBI traumatic brain injury*Fisher’s test for categorical variables, Wilcoxon’s test for continuous variables$ Please see supplement 1 for more details Patient characteristics Numbers are count (percentage) or median [IQR] APACHE acute physiology and chronic health evaluation; Combi-drain, external ventricular drainage draining cerebrospinal fluid and measuring intracranial pressure simultaneously, GCS Glasgow Coma Scale, ICH Intracerebral haemorrhage, ICU intensive care unit, MMNM multimodal neuromonitoring, SAH subarachnoid haemorrhage, TBI traumatic brain injury *Fisher’s test for categorical variables, Wilcoxon’s test for continuous variables $ Please see supplement 1 for more details Table 2MMNM modalities used (N = 223)ModalitiesN (%)PbtO2 and ICP*6 (2.7%)PbtO2 and ICP33 (15%)PbtO2 and CMD96 (43%)PbtO2, ICP and CMD71 (32%)PbtO2, ICP, Tbt and CMD17 (7.6%)CMD cerebral microdialysis, ICP intracranial pressure, PbtO2 brain tissue oxygen tension, Tbt Brain temperature*In 6 patients, no ICP catheter was inserted through the double bolt (IM2SEU) due to surgical difficulties, and contrary to the intention. Thus, ICP was measured using an EVD combi-drain or a tunnelled ICP catheter, and only the PbtO2 MMNM modalities used (N = 223) CMD cerebral microdialysis, ICP intracranial pressure, PbtO2 brain tissue oxygen tension, Tbt Brain temperature *In 6 patients, no ICP catheter was inserted through the double bolt (IM2SEU) due to surgical difficulties, and contrary to the intention. Thus, ICP was measured using an EVD combi-drain or a tunnelled ICP catheter, and only the PbtO2 Table 3Anatomical location of MMNM equipment by clinical diagnosis (N = 223)*ConditionHemisphere locationN (%) of patientsSAH (non-midline aneurysm)Contralateral hemisphere44 (20%)Ipsilateral hemisphere44 (20%)SAH (midline aneurysm)Right hemisphere8 (3.5%)Left hemisphere3 (1.3%)Aneurysm not foundRight hemisphere2 (0.9%)Left hemisphere1 (0.4%)TBIMost injured hemisphere41 (18%)Contralateral hemisphere76 (34%)ICHContralateral hemisphere2 (0.9%)Ipsilateral hemisphere2 (0.9%)*At first placement. Subsequent insertion generally on same side, except for three patients with TBIICH intracerebral haemorrhage, SAH subarachnoid haemorrhage, TBI traumatic brain injury Anatomical location of MMNM equipment by clinical diagnosis (N = 223)* *At first placement. Subsequent insertion generally on same side, except for three patients with TBI ICH intracerebral haemorrhage, SAH subarachnoid haemorrhage, TBI traumatic brain injury MMNM-related complications Table 4 shows an overview of surgical complications after primary insertion of MMNM. Haemorrhage was observed in 25 patients (13%), comprising 18 patients (9%) with subdural haematoma (median thickness, 0.4 cm (IQR 0.3–0.5)) and 9 (5%) with intracerebral haemorrhage (Grade I = 5 (2.2%), Grade II = 0, and Grade III = 4 (1.7%), maximum size 3.3 ml); none required surgical treatment. In 2 patients, haemorrhage was present both subdurally and in the parenchyma. No haemorrhage was observed after catheter reinsertion. Table 4MMNM surgery related complications after primary insertionComplicationsAll probesIntegra LifeSciencesDouble bolt (PbtO2 only*)Double boltTriple boltRaumedic$(N = 196)(N = 6)(N = 114)(N = 62)(N = 14)Haemorrhage25 (13%)3 (50%)12 (101%)9 (15%)1 (7.1%) ICH9 (5%)1 (17%)5 (4.4%)2 (3.2%)1 (7.1%) SDH18 (9%)2 (33%)8 (7.0%)8 (12.9%)0 (0%)Volume of ICH (ml)0.6 [0.2–2.2]2.7 [2.7–2.7]0.2 [0.2–0.6]1.1 [0.6–1.6]1.6 [1.6–1.6]Width of SDH (cm)0.4 [0.3–0.5]0.9 [0.9–0.9]0.3 [0.3–0.4]0.4 [0.3–0.5]-Suboptimal catheter placement70 (36%)4 (67%)38 (33%)24 (39%)4 (29%)Bone fragments25 (13%)2 (33%)12 (11%)6 (10%)5 (36%)Pneumocephalus71 (36%)2 (33%)40 (35%)26 (42%)3 (21%)Numbers are count (percentage) or median (IQR)ICH intracerebral haemorrhage, IQR inter-quartile range, MMNM multimodal neuromonitoring, SDH subdural haemorrhage*In 6 patients, no ICP catheter was inserted through the double bolt (IM2SEU) due to surgical difficulties, and contrary to the intention. Thus, ICP was measured using an EVD combi-drain or a tunnelled ICP catheter, and only the PbtO2$ Neurovent-PTO 2L Bolt Table 4MMNM surgery related complications after primary insertionComplicationsAll probesIntegra LifeSciencesDouble bolt (PbtO2 only*)Double boltTriple boltRaumedic$(N = 196)(N = 6)(N = 114)(N = 62)(N = 14)Haemorrhage25 (13%)3 (50%)12 (101%)9 (15%)1 (7.1%) ICH9 (5%)1 (17%)5 (4.4%)2 (3.2%)1 (7.1%) SDH18 (9%)2 (33%)8 (7.0%)8 (12.9%)0 (0%)Volume of ICH (ml)0.6 [0.2–2.2]2.7 [2.7–2.7]0.2 [0.2–0.6]1.1 [0.6–1.6]1.6 [1.6–1.6]Width of SDH (cm)0.4 [0.3–0.5]0.9 [0.9–0.9]0.3 [0.3–0.4]0.4 [0.3–0.5]-Suboptimal catheter placement70 (36%)4 (67%)38 (33%)24 (39%)4 (29%)Bone fragments25 (13%)2 (33%)12 (11%)6 (10%)5 (36%)Pneumocephalus71 (36%)2 (33%)40 (35%)26 (42%)3 (21%)Numbers are count (percentage) or median (IQR)ICH intracerebral haemorrhage, IQR inter-quartile range, MMNM multimodal neuromonitoring, SDH subdural haemorrhage*In 6 patients, no ICP catheter was inserted through the double bolt (IM2SEU) due to surgical difficulties, and contrary to the intention. Thus, ICP was measured using an EVD combi-drain or a tunnelled ICP catheter, and only the PbtO2$ Neurovent-PTO 2L Bolt MMNM surgery related complications after primary insertion Numbers are count (percentage) or median (IQR) ICH intracerebral haemorrhage, IQR inter-quartile range, MMNM multimodal neuromonitoring, SDH subdural haemorrhage *In 6 patients, no ICP catheter was inserted through the double bolt (IM2SEU) due to surgical difficulties, and contrary to the intention. Thus, ICP was measured using an EVD combi-drain or a tunnelled ICP catheter, and only the PbtO2 $ Neurovent-PTO 2L Bolt One or more catheters were found to be suboptimally located after first insertion in 70 patients (36%) (sub-categories are found in Supplementary Table 3). Eight out of 37 (22%) and 2 out of 6 (33%) patients, respectively, had suboptimally located catheters after the first and second reinsertion. Malfunction of one or more catheters was one of the most common complications with 26 of 62 (42%) verified cases; however, a suboptimal located catheter was not related to malfunction (OR for malfunction, suboptimal vs. optimal location, 0.64 (95% CI: 0.36–1.16), p = 0.14). Dislodgement occurred in 19 cases. Seventy-one patients (36%) had MMNM-related pneumocephalus consisting mostly of small air bubbles along the catheters (66%) or extra-axially (34%) (OR for malfunction given pneumocephalus, 0.95 (95% CI: 0.53–1.72), p = 0.87). Only one patient had pneumocephalus on CT after a reinsertion. Bone fragments were detected in 25 patients (13%) after primary insertion (OR for malfunction given bone fragments, 3.7 (95% CI: 0.76–14.5), p = 0.19). One patient had bone fragments after a reinsertion. Forty-five patients (20%) were treated with antibiotics for suspected intracranial infection. In one patient (0.4%), Staphylococcus haemolyticus was cultured from a PbtO2 monitoring catheter; the remaining infections were thought to be ventriculitis associated with the presence of an EVD, and the indication for antibiotic treatment was not given as an MMNM-related infection in the medical chart. No skin infection occurred in relation to MMNM. In total, at least one MMNM-related complication was found in 165 patients (74%) after insertion, at a total of 272 complications (Fig. 1). Subgroup comparison can be found in Supplementary Table 4 to explore potential clinical risk factors of different types of complications. These analyses showed no statistically significant differences in MMNM‑related complications between bedside and operating room insertion of MMNM.Fig. 1Distribution of the number of complications per patient Distribution of the number of complications per patient Complications after MMNM removal After removal of MMNM, 162 patients (73%) underwent CT or MRI, of whom 79 patients (49%) had at least one new MMNM-related complication compared to after insertion. Thus, 11 patients (6.8%) developed a parenchymal haemorrhage related to MMNM (median volume, of 0.09 (IQR: 0.003–0.18) ml). One patient had an intracerebral haemorrhage (28 ml) which was surgically evacuated. Three patients with a parenchymal haemorrhage did not have a CT scan after MMNM insertion. One patient developed an SDH after removal. Lastly, 66 patients (40%) had bone fragments, of whom 12 did not have a CT scan after insertion, and 13 patients (8%) had some degree of pneumocephalus, of whom 3 did not have a CT scan after insertion. Association between MMNM complications and patient outcomes No difference in 6 months functional outcome or mortality was found between patients with and those without complications related to MMNM (Table 1). Unfavourable functional outcome on the mRS scale appeared in 68% of patients with MMNM-related complications compared to 67% of those without, and 76% of patients with MMNM-related complications compared to 71% of those without had unfavourable GOSE outcome. Six months mortality was 32% and 26%, for patients with and those without MMNM-related complications, respectively. Discussion MMNM is standard of care in many Neuro-ICUs for patients with severe acute brain injury and is generally considered safe with complication rates consistent with other intracranial devices used for these types of patients. Using our definition of MMNM-related complications, we found a high overall complication rate, with 165 patients (74%) having at least one complication, most of which were attributed to malfunction, suboptimally located catheters, and pneumocephalus. However, the vast majority were trivial, did not require interventions, and did not affect patient treatment; only one patient required surgical treatment due to a haematoma occurring after removal of MMNM. Importantly, MMNM-related complications were not associated with unfavourable functional outcome or mortality after 6 months. Haemorrhage We observed 25 patients (13%) with haemorrhage after MMNM insertion, which is within the previously reported range from 0–40% [5, 10]. Notably, this is also within the range of EVD-related haemorrhage ranging from 0–41% [16, 19]. Few publications report the need for surgical treatment of MMNM-related haemorrhage [1, 3, 5, 15]. In our cohort, one patient required surgery for removal of an intracerebral haematoma that developed after the MMNM equipment was removed. This corresponds to previous reports, where most haemorrhages were clinically nonsignificant compared to the primary brain injury and did not require intervention [5, 15]. Suboptimally located catheters and malfunction In our cohort, one of every three catheters were suboptimally located, which is high compared with other studies, with a reported rate ranging from 3–14% [1, 3, 10, 21]. We pre-specified a rather wide definition of suboptimally located catheters with nine sub-categories, many of which have not previously been reported elsewhere in the literature. This may have contributed to the lack of an association with malfunction; thus, despite our findings we continue to believe that careful positioning of MMNM devices by trained neurosurgeons is important to obtain clinical benefit. Malfunction was reported for one out of four catheters, but close to 60% of these were non-verified, meaning that it was not validated by an MMNM-trained clinician. For comparison, the literature reports rates of malfunction between 5–43% [1, 3, 21]. As a limitation of our retrospective study design, the specific causes of catheter malfunctioning were difficult to categorise. Besides functional testing of the equipment, distinguishing between correctly functioning and malfunctioning equipment mainly depends on the training and experience of the clinician. Bailey et al. [3] reported that 5% of all PbtO2 catheters were replaced on suspicion of malfunctioning, but that 54% of these were confirmed to be functioning normally, which corroborates the need for adequate training of the staff. Bone fragments and pneumocephalus Few reports have reported the prevalence of pneumocephalus and bone fragments in brain parenchyma after insertion or removal of MMNM, let alone their clinical consequences [1, 5, 10, 11]. When reported, they are usually categorised as clinically asymptomatic. In our study, none of these phenomena were documented in the patient records or treated, and no clinical consequences were observed. The long-term clinical significance of bone fragments and pneumocephalus, however, remains unclear. Bone fragments were more prevalent after MMNM removal, probably due to obscuring artefacts while bolts and catheters were still in place. Infection In our study cohort, one patient was treated for an MMNM-related infection; others have reported prevalences between 0–5% [1, 3, 10, 21]. Although it can be challenging to determine whether central nervous system (CNS) infections originate from MMNM or EVD, the risk of MMNM-related infections seems low compared to, e.g. that of EVD-related CNS infections, the prevalence of which ranges from 0–28% [16]. Patients with severe acute brain injury, which inherently has a high morbidity and mortality, may benefit from MMNM as it provides crucial information on the ongoing pathophysiology and potential risk of secondary brain injury [7, 12, 14, 17, 18]. MMNM gives clinicians the opportunity to personalise treatment according to the patient’s current condition. Although results from the BOOST-3 [6] and the BONANZA [23] trials in patients with TBI are still awaited, published studies suggest that combined PbtO2- and ICP- guided therapy may improve outcome [12, 17, 18] in those with high ICP [20]. Strengths and limitations The major strength of this study was the large cohort of patients with severe acute brain injury, both TBI and non-TBI. Additionally, two authors independently assessed CT and MRI scans, and two separate authors independently assessed functional outcome. However, this study also has limitations. Because it was a single-centre retrospective study, the external validity remains limited; though, the findings mostly reflected previous studies. A surgical complication scale, such as the Clavien Dindo Scale of Surgical Complications [8] could have been useful for comparison and external validation. Nevertheless, this scale is intended for more general elective surgery, and to our knowledge no similar scale has been devised for emergency surgery. Next, the lack of formal radiology training of the authors in reviewing head CTs or MRIs is a limitation as this could lead to over- or underestimation of the frequency of surgical complications. It would have been more appropriate, but also resource-demanding to have a radiologist review the head CTs and MRIs. Another limitation of the study is that a postoperative CT scan was not performed systematically as standard of care, but only on clinical indication, which could lead to underestimation of the surgical complications found on the CT scans after insertion. Lastly, we did not assess the level of experience of neurosurgeons performing the procedure in the analysis, nor did we compare complications occurring during daytime or night shifts. Conclusion We found a high rate of MMNM-related complications. Most were believed to be MMNM-related findings rather than true complications, as they did not cause clinical concern, with very few requiring interventions. We found no association between MMNM-related complications and functional outcome or mortality after 6 months. MMNM appears generally safe to patients with severe acute brain injury. However, greater knowledge of MMNM-guided therapy on patient outcome is required to support the current use of MMNM. Supplementary Information Below is the link to the electronic supplementary material.ESM 1Additional supporting information is available online in the Supplementary Materials. Supplementary Table 1: Subgroups of comorbidities and Neuro-ICU infection. Supplementary 2: Severity scoring of patients with SAH. Supplementary Table 3: Nine subgroups of suboptimal located catheters following the initial insertion of MMNM. Supplementary 4: MMNM-related complications categorized for comparative analysis to identify potential clinical risk factors associated with different complication types. (DOCX 39.5 KB) Additional supporting information is available online in the Supplementary Materials. Supplementary Table 1: Subgroups of comorbidities and Neuro-ICU infection. Supplementary 2: Severity scoring of patients with SAH. Supplementary Table 3: Nine subgroups of suboptimal located catheters following the initial insertion of MMNM. Supplementary 4: MMNM-related complications categorized for comparative analysis to identify potential clinical risk factors associated with different complication types. (DOCX 39.5 KB)