Timing Is Critical in Determining the Association Between Delirium and S100 Calcium‐Binding Protein B

The pathophysiology of delirium is poorly understood, but the higher incidence of dementia after delirium has led to the hypothesis that some individuals with delirium might have irreversible brain damage. The protein S100 calcium-binding protein B (S100B) is a possible marker of this brain damage.1 Astrocytes and, to a lesser extent, extraneuronal tissues express S100B. At nanomolar concentrations, S100B exerts trophic effects on neuronal tissue; at micromolar concentrations, its effects are merely toxic.2 Previous research has shown that S100B serum levels are high after trauma or surgery, during inflammation, and in individuals with Alzheimer's disease.1, 2 High S100B levels have been found in the serum of individuals with delirium in different settings.3-7 A disadvantage of measuring S100B levels in serum is that they do not necessarily reflect brain damage, because extraneuronal sources can also contribute to high S100B levels.8 Studying cerebrospinal fluid (CSF) levels of S100B can bypass this problem. Higher CSF levels of S100B were found in eight individuals with hip fracture with delirium at the time of CSF sampling than in 38 individuals who were not (yet) delirious.9 CSF S100B levels of individuals with and without delirium were studied. It was hypothesized that individuals with delirium would have higher levels of S100B than those without delirium. Sixty-six individuals aged 65–97 who were acutely admitted to a teaching hospital in Amsterdam for surgical treatment of hip fracture were included. Demographic data and medical history were recorded. A caregiver completed the Informant Questionnaire on Cognitive Decline Short Form. Cognitive impairment was defined as a score of 3.4 or higher or documented dementia diagnosis. The presence of delirium in nursing and medical records was assessed using Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, criteria. CSF samples were collected preoperatively during cannulation for spinal anesthesia, before administration of anesthesia. Samples were centrifuged and stored at −80°C until required. S100B concentration was measured using an enzyme-linked immunosorbent assay against human S100B (Millipore, Darmstadt, Germany). The lowest detection limit of this assay is 2.7 pg/mL; levels below this limit were set at half this value. Difference in S100B levels between groups was assessed using the Mann-Whitney U test. The hospital's medical ethical committee approved the study. Fifteen participants developed delirium during hospitalization, two of whom already had delirium before surgery. There was no difference in baseline characteristics between participants with and without delirium. The mean age of participants with delirium was 86.1 ± 8.4 and of those without delirium was 82.9 ± 8.0 (P = .18). Four participants with delirium (26.7%) and 11 without (21.6%) had preexisting cognitive impairment (P = .68). Preoperative S100B levels did not differ between participants with (median 1,053 pg/mL, interquartile range (IQR) 601–1,178 pg/mL)) and without (median 862 pg/mL, IQR 701–1,156 pg/mL) delirium (P = .76). The two participants with preoperative delirium had S100B levels of 1,052pg/mL and 2,258pg/mL—above the median levels of participants with postoperative (median 848 pg/mL, IQR 595–1,177 pg/mL) or no (median 862 pg/mL, IQR 701–1,156 pg/mL) delirium (Figure 1). The present study in 66 elderly adults with hip fracture showed no difference in preoperative CSF S100B levels between those with and without delirium after hip fracture. The higher S100B levels found in two participants with delirium at the time of CSF sampling suggests that timing is important in determining the association between delirium and S100B levels. This is in accordance with research that found that S100B was higher if measured during an episode of delirium than before or after.4, 9 Only two studies have previously investigated an association between CSF S100B levels and delirium. The high CSF S100B levels in individuals with delirium found in one must be interpreted with caution because of the small sample size.9 The other found no difference in CSF S100B levels between 20 individuals with long-lasting delirium and 20 with probable Alzheimer's disease.10 Because of lack of a comparison group without delirium and Alzheimer's disease, interpretation of these data is difficult. There is a clear demand for more research on the association between delirium and S100B to establish its role as an end-stage disease marker in delirium. A high-priority topic is the release pattern of S100B in CSF during delirium and the interaction with inflammatory markers. More knowledge on the relationship between S100B levels during delirium and long-term cognitive and functional outcomes is also of interest, because this might shed light on the assumed relationship between delirium and subsequent cognitive and functional decline. We would like to thank Prof. Dr. Adriaan Honig for psychiatric consultations, Dr. Ineke van der Waart for CSF sample collection, and Dr. Manouk Backes and Dr. Joanne Sierink for their support in collecting data. This work was partially supported by ZonMW (Westhoff and van de Beek, TOP Grant 40–00812–98–10017, ZonMw VIDI), and the European Research Council (van de Beek, starting grant). Author Contributions: De Rooij, Van Munster: study design. Vellekoop: design and implementation of local study protocol. Scholtens, Vrouenraets: data collection. Scholtens, Westhoff: laboratory testing. van de Beek: supervision of laboratory testing. Beishuizen, Scholtens, De Rooij, Van Munster: data analysis. Beishuizen, De Rooij, Van Munster: drafting manuscript. All authors read and approved the final version of the manuscript. Sponsor's Role: None.