Hourly analysis of cerebrospinal fluid glucose shows large diurnal fluctuations

Abstract

Cerebrospinal fluid analysis is important in the diagnostics of many neurological disorders. Since the influence of food intake on the cerebrospinal fluid glucose concentration and the cerebrospinal fluid/plasma glucose ratio is largely unknown, we studied fluctuations in these parameters in healthy adult volunteers during a period of 36 h. Our observations show large physiological fluctuations of cerebrospinal fluid glucose and the cerebrospinal fluid/plasma glucose ratio, and their relation to food intake. These findings provide novel insights into the physiology of cerebral processes dependent on glucose levels such as energy formation (e.g. glycolysis), enzymatic reactions (e.g. glycosylation), and non-enzymatic reactions (e.g. advanced endproduct glycation).

Introduction

Cerebrospinal fluid (CSF) analysis is important for the diagnostic work-up of many neurologic disorders. The CSF glucose concentration and the CSF/plasma glucose ratio are helpful measurements for the differential diagnosis of acute and often treatable disorders. In meningitis, for example, a low CSF glucose and low CSF/plasma glucose ratio (<0.44) suggest a bacterial rather than a viral cause.¹ The CSF/plasma glucose ratio normally ranges from 0.5 to 0.8.^2,3

It is assumed that a steady state between blood and CSF glucose is reached 4 to 6 h after a meal and therefore a lumbar puncture should ideally be performed in the fasting state. This assumption is based on a study with a single intravenous bolus of glucose.⁴ No reports are available on the variation of blood and CSF glucose over a 24-h period in response to meals, and, thus, on the interpretation of CSF results in the non-fasting state. Furthermore, since glucose is the most important source of fuel for the brain, more detailed insights in the regulation and fluctuations of cerebral glucose levels may provide more insights into the fluctuations of processes that depend on levels of cerebral glucose such as energy formation (glycolysis, citric acid cycle), enzymatic glycosylation reactions, or adverse reactions associated with high glucose, such as non-enzymatic advanced glycation endproduct formation.

Methods

We have studied the diurnal fluctuations of the CSF glucose concentration and the CSF/plasma glucose ratio under physiological conditions in five healthy volunteers aged 59–85 years. Healthy participants were initially included in the study for biomarkers on Alzheimer’s disease. In this study, we excluded the CSF samples of one of these participants, a male aged 64 years, because of an adverse event during the study (deep vein thrombosis with pulmonary embolus), which did not make this participant a suitable candidate for analysis of CSF glucose and lactate in the normal, physiological state. The study was approved by the local institutional review board (Commissie Mensgebonden Onderzoek Arnhem-Nijmegen), and all participants gave written informed consent to receive a spinal intrathecal catheter for 36 h. The study was conducted in accordance with the provisions of the Helsinki Declaration. Paired serum and CSF samples were collected every hour. For a detailed description of the methods, we refer to a previous publication.⁵ Glucose was measured with the hexokinase method, using a Cobas Mira analyzer.³

Results

Plasma glucose varied between 4.4 and 9.6 mM (mean 6.0; SD 1.1), CSF glucose between 2.8 and 4.4 mM (mean 3.4; SD 0.34), and the CSF/plasma glucose ratio between 0.30 and 0.89 (mean 0.59; SD 0.10). CSF glucose and the CSF/plasma glucose ratio showed large fluctuations within the same individual, even exceeding upper and lower reference ranges (Figure 1).³ In all participants, the CSF/plasma glucose ratio was high just before meals and low during the first hours thereafter. A high CSF/plasma ratio was found in the late evening, followed by a decrease during the night. The CSF/plasma ratio stabilized after 4 to 6 h of fasting during the night.

Figure 1.

(a) Typical diurnal fluctuation of CSF glucose (mM) (black line) and blood glucose (mM) (grey line) in relation to meals in participant 1. (b) Diurnal fluctuation of the CSF/blood glucose ratio in participant 1 (black line) and participants 2–5 (gray lines). The dotted line represents the mean. Red bars indicate mealtime in participant 1: breakfast and dinner consisted of bread (45–60 g of carbohydrates in a meal); lunch consisted of a hot meal with meat, potatoes, and vegetables (80–100 g of carbohydrates in a meal). Gray bar indicates nighttime.

Discussion

This study demonstrates the physiological fluctuations of CSF glucose and the CSF/plasma glucose ratio, and their relation to food intake. The data clearly show that there is a period within the first hours after a meal in which CSF glucose is low, but in which plasma glucose has already increased, resulting in a markedly low CSF/plasma glucose ratio. This reflects the fact that plasma glucose changes fast and CSF glucose only follows slowly after food intake. The result is a low CSF/plasma glucose ratio that gives the impression of a pathological low ratio instead of a physiological phenomenon. A diagnostic lumbar puncture during the first hours after a meal therefore may lead to errors in interpretation. In the late evening, a high CSF/plasma glucose ratio is followed by a decrease during the night that is not preceded by a meal, which might suggest that other factors besides food intake, like circadian hormone rhythms, influence the CSF/plasma glucose ratio as well.

Based on experimental data from the 1960s, it is believed that the equilibration of plasma to CSF glucose takes about 2 to 4 h.⁴ During these experiments, five participants received a single intravenous injection of 50% glucose followed by plasma and CSF glucose analysis during a period of 6 h. In contrast to these previous experiments, our study provides important information on the fluctuation of the CSF/blood glucose ratio during a full day under physiological circumstances. An alternative explanation to the effect of food intake on cerebral glucose levels is the observation that the rate of CSF production also may be dependent on the time of day.⁶ In a small study, it was demonstrated that CSF flow, as a measure of CSF production rate, could be 80% higher at night when compared to end of the day levels.

Our data suggest that the cerebral glucose concentrations vary considerably over the day. Literature findings suggest that these fluctuating glucose concentrations may differentially affect several biochemical processes in the brain. For example, the degree by which proteins like tau become modified by either O-Gluc-NAc addition or hyperfosforylation, which are mutually exclusive processes, is dependent on cerebral glucose levels.⁷ High glucose levels may also enhance the level of irreversible protein modification by advanced glycation end products.⁸

In conclusion, the time since the last meal should be considered for correct interpretation of CSF glucose results. Our findings emphasize that a diagnostic lumbar puncture ideally should be performed after at least 4 h of fasting. Finally, the naturally occurring fluctuations in cerebral glucose levels may be important in the regulation of several biochemical mechanisms that have been associated with Alzheimer's disease.

Funding

Dr. Leen was supported by a grant from NWO, ZonMW, (“AGIKO-stipendium”; grant number 92003529). Dr. Verbeek was supported by a grant from ZonMW (CAVIA project; nr. 733050202). The CAVIA project is part of ‘Memorabel’, the research and innovation programme for dementia, as part of the Dutch national ‘Deltaplan for Dementia’: zonmw.nl/dementiaresearch”. The CAVIA project (www.caviaproject.nl) is a consortium of Radboudumc, LUMC, Erasmus MC, VUmc, ADX Neurosciences, Philips Healthcare, Stony Brook University and Massachusetts General Hospital.

Declaration of conflicting interests

M.M. Verbeek served on an advisory board for Roche. W.G. Leen has received a grant for performing research on the subject of glucose transport into the brain (NWO, ZonMW, “AGIKO-stipendium”; grant number 92003529). M.A. Willemsen reports no disclosures. D. Slats reports no disclosures. J.A. Claassen reports no disclosures. All coauthors have seen and agree with the contents of this manuscript. All authors declare that there are no conflicts of interest. The corresponding author takes full responsibility for the data and interpretation. The corresponding author has full access to all of the data.

Authors’ contributions

MM Verbeek: design of the study, supervision of experimental data analysis, supervision of the study, data analysis, revising the article critically for important intellectual content, and final approval.

WG Leen: data analysis, drafting the article and final approval of the article, obtained funding.

MA Willemsen: design of the study, supervision of the study, data analysis, revising article critically for important intellectual content and final approval, obtained funding.

D. Slats: execution of clinical study and collection of data, revising article critically for important intellectual content, and final approval.

J.A. Claassen: design of the study, supervision of the study, data analysis, revising article critically for important intellectual content, and final approval.