Body fluids

Formation and physiologic role

In health, CSF is a clear, aqueous liquid that surrounds the brain and the spinal cord, providing buoyant physical support for the brain. It also has a role in intracerebral transport of biomolecules, removal of CNS waste metabolites, maintenance of constant intracranial pressure, and defense against invasion by pathogens. Examination of CSF may provide critically important diagnostic information in a number of infectious and noninfectious medical conditions.

CSF is produced in the choroid plexus, an intricate network of capillaries, epithelial cells, and interstitial connective tissue found in the lateral, third, and fourth ventricles. CSF circulates in the subarachnoid space, where flow is primarily back toward the cerebral hemispheres to the parasagittal region where reabsorption occurs ( Fig. 45.5 ). A small amount of flow is also directed down the brainstem and spinal cord, ultimately returning to the cerebral subarachnoid space, where CSF is reabsorbed in the arachnoid villi. Functioning as valves, the arachnoid villi are located along the superior sagittal and intracranial venous sinuses and around the spinal nerve roots.

CSF is generated in the choroid plexus through processes including ultrafiltration and active transport ( Fig. 45.6 ). A Na ⁺ /K ⁺ -ATPase in the choroid plexus actively transports Na ⁺ into the CSF, causing free water to follow. Choroid epithelial cells contain carbonic anhydrase, and the counter-transport of chlorine and bicarbonate is integral to CSF formation. Other metabolites, including vitamins, are secreted into the CSF by specific transporters, whereas physiologic metabolites such as hydrogen, carbon dioxide, and ammonia may pass by simple diffusion. Proteins are transported from serum to CSF by pinocytosis or specific carriers.

The total volume of CSF in adults is 150 mL, generated at an average rate of 20 mL/hr, thus CSF is in a dynamic circulation that is replaced every 5 to 7 hours. Normal CSF pressure measured with a manometer in a patient lying in the lateral decubitus position with the legs extended is between 60 and 250 mmH ₂ O, although some authorities consider the upper limit of normal CSF pressure to be 200 mmH ₂ O. Processes such as infection, bleeding, or a tumor can cause intracranial hypertension. Slow-growing masses such as abscesses or tumors may allow time for compensation between CSF secretion and absorption to occur, and a rise in CSF pressure may not necessarily be observed.

CSF is normally clear and colorless. The fluid will appear slightly turbid with the presence of more than 200 × 10 ⁶ WBCs/L. Up to 5 × 10 ⁶ WBCs/L (lymphocytes and monocytes) are considered normal in adults and up to 20×10 ⁶ WBCs/L (predominantly polymorphs) in neonates. An increased CSF WBC count does not necessarily indicate an infection because increases can occur in both infectious and noninfectious inflammatory states. Erythrocytes are not normally present in CSF and may be introduced after trauma to a blood vessel during the procedure (“traumatic tap”) or as a result of SAH (see later).

Evaluation for meningitis (glucose, lactate, lactate dehydrogenase, C-reactive protein, adenosine deaminase)

Glucose concentration in CSF is derived from plasma glucose by both facilitated transport and simple diffusion. Glucose is used by cells lining the ventricular cavities and subarachnoid spaces and, as a result, it normally takes 2 to 4 hours for plasma glucose to equilibrate with CSF glucose. The normal ratio of CSF to serum glucose is approximately 0.6 to 0.8; ventricular CSF has a higher glucose concentration than CSF in the lumbar space by 6 to 18 mg/dL (0.33 to 1.0 mmol/L).

For reporting, CSF glucose must be interpreted in association with a plasma glucose sample collected at the same time. Abnormally low CSF glucose concentrations can occur in bacterial meningitis and also mycobacterial, mycoplasmal, and fungal CNS infections. Using a CSF to serum glucose ratio of less than 0.4, an 80% sensitivity and 98% specificity was found for distinguishing bacterial ( n = 119) versus aseptic cases ( n = 97) of meningitis. During recovery from meningitis, the CSF glucose concentration tends to normalize more rapidly than do the CSF cell count and protein concentration. The CSF glucose concentration is typically normal during most viral CNS infections, although exceptions may occur in patients with meningoencephalitis secondary to mumps, enteroviruses, herpes simplex, and herpes zoster viruses. It should be noted that decreased CSF glucose may result from changes in the physiologic function of the choroid epithelium and from consumption by bacterial pathogens and leukocytes; thus altered CSF glucose may reflect sequelae of disease rather than microbial invasion. Low CSF glucose concentrations can be observed in noninfectious processes, including malignant disease infiltrating the meninges, SAH, and CNS sarcoidosis. However, CSF glucose concentrations less than 18 mg/dL (1.0 mmol/L) are strongly predictive of bacterial meningitis. The ratio of CSF to serum glucose has limited utility in neonates and in patients with severe hyperglycemia. CSF glucose concentrations rarely exceed 300 mg/dL (16.7 mmol/L), even in patients with severe hyperglycemia. A low concentration of glucose in the CSF in combination with a decreased ratio of CSF to blood glucose may also be encountered in the inherited condition of GLUT 1 transporter deficiency, which may manifest with intellectual disability, intractable seizures, and motor impairment. In these cases, a low to normal concentration of CSF lactate discriminates GLUT1 deficiency syndrome from other causes of low CSF glucose, and this approach has been advocated as a triage before progressing to more specialized mutation analysis in suspected cases.

CSF lactate is predominantly produced by CNS glycolysis and is independent of serum lactate. Increased CSF lactate concentrations are related to increased cerebral glycolysis secondary to hypoxia, ischemia, seizures, SAH, or meningitis. Determination of the CSF lactate concentration has been proposed as a test to differentiate bacterial from viral meningitis. Two meta-analyses including 56 studies and 3577 patients concluded that the diagnostic accuracy of CSF lactate was superior to that of CSF WBC count, glucose, and protein concentration in differentiating bacterial from aseptic meningitis, ^, although its sensitivity was lower in patients who had received antimicrobial treatment before LP. CSF lactate may be increased in patients with other CNS diseases. Normal CSF lactate is 1.1 to 2.3 mmol/L (10 to 21 mg/dL), and concentrations greater than 4.2 mmol/L (40 mg/dL) are suggestive of bacterial meningitis, while values below this value are suggestive of viral meningitis. In a study of 53 cases of diagnosed meningitis, 24 of 25 patients with bacterial meningitis (culture or positive Gram stain) initially presented with a CSF lactate greater 4.2 mmol/L (37.8 mg/dL). None of the 28 patients with presumed viral meningitis (supported by negative cultures and spontaneous resolution without antibiotics) showed CSF lactate concentrations above 4.2 mmol/L (37.8 mg/dL). In addition, the ratio of CSF to blood glucose of 0.4 in this group allowed the differentiation of patients into bacterial and aseptic cases with almost equal sensitivity (96% by lactate versus 91% by glucose ratio). Decreasing CSF lactate concentrations may be interpreted as evidence of effective treatment of the infection.

LD is an intracellular enzyme that catalyzes the final step in anaerobic glycolysis. Release of LD into the CSF occurs as a result of cell death. It is therefore a nonspecific marker of cell necrosis and may be raised in a number of pathologic conditions, including meningitis, with a markedly higher rise in bacterial versus viral meningitis. The CSF LD reference interval is 0 to 23.5 U/L. CSF LD is also increased in primary brain tumors and metastases, particularly those involving the meningeal membrane, and hematologic metastases, including leukemic infiltration. It is also raised in cases with increased intracranial pressure (hydrocephalus), seizures, hypoxia, and cerebral ischemia. Determination of LD isoenzymes in the CSF does not confer any additional clinical utility.

C-reactive protein (CRP) has been proposed as another analyte that may offer utility in discrimination of bacterial and viral meningitis. The majority of CSF CRP is derived from serum, where it is produced by the liver as an acute-phase reactant in response to several disease states. Intrathecal synthesis of CRP is minimal. Many studies have been undertaken with contrasting findings, and the overall conclusion is that CRP is insensitive, nonspecific, and too dependent on a functioning host inflammatory response to make it diagnostically useful for distinguishing septic from aseptic meningitis. It does not confer any diagnostic advantage over the more established markers of CSF cell count, protein, glucose, and lactate. Similar conclusions have also been reached for CSF cytokines, including tumor necrosis factor-alpha (TNF-α), interleukin 1 (IL-1), and interleukin 6 (IL-6), where the concentrations present depend on the ability of the host to mount an immune response.

Adenosine deaminase (ADA) is an enzyme present in leukocytes, released in response to increased inflammatory activity after CNS infection and has been proposed as a marker to distinguish tuberculous (TB) meningitis from other forms of meningitis. It was found that CSF ADA was increased in 64% of TB and bacterial meningitis cases, a sensitivity that conferred no advantage over protein or glucose measurement. Furthermore, patients with sepsis demonstrated a wide range of concentrations. Other studies have shown similar findings, and it is concluded that CSF ADA does not confer any diagnostic advantage over CSF glucose or protein.

Evaluation of demyelinating diseases (cerebrospinal fluid index, oligoclonal banding)

Under normal circumstances, immunoglobulins are almost totally excluded from the CSF, with a ratio of blood to CSF IgG normally 500:1 or greater. The detection of restricted clones of immunoglobulin synthesis in the CSF, termed oligoclonal bands, may occur in any disorder that disrupts the blood–brain barrier, including systemic inflammation. In demyelinating disorders, however, the detection of strictly intrathecal IgG synthesis is an important diagnostic criterion, although it should be complementary to clinical assessment and radiologic imaging. In these conditions, infiltration with B lymphocytes leads to the release of immunoglobulins, primarily IgG. The CSF IgG concentration may be increased because of increased plasma concentration, impaired barrier function, or increased intrathecal synthesis. Methods for the determination of increased intrathecal synthesis include detection of CSF oligoclonal banding visually by electrophoresis or isoelectric focusing. For further discussion about these techniques refer to Chapter 18 . Another approach is to calculate the CSF IgG index or synthesis rate, which is often used in a complementary fashion.

The CSF IgG index is calculated as:

CSF concentrations in mg/L; serum concentrations in g/L.

The reference interval for the CSF IgG index is 0.3 to 0.7, and ratios greater than 0.7 indicate increased intrathecal synthesis, as seen in more than 80% of cases of multiple sclerosis, although this can also be seen in other demyelinating conditions. ^, It should be noted, however, that in certain conditions in which the integrity of the blood–brain barrier is compromised, including SAH and traumatic LP, the IgG index can be falsely increased. Interpretation of CSF IgG concentrations and derived indices is much more reliable when the albumin index (see “Proteins”), a marker of the integrity of the blood–brain barrier, is also normal.

The rate of intrathecal IgG synthesis can be estimated by the empirically derived Tourtellotte’s formula, and a synthesis rate of greater than 8 mg/day is found in most cases of multiple sclerosis. It is a more complex formula with several constants and probably provides no additional information over the IgG index.

For determination of CSF oligoclonal banding by electrophoresis or isoelectric focusing, it is essential that a serum sample is analyzed concurrently so that intrathecal synthesis may be distinguished from banding that originated in the serum and diffused across the blood-CSF barrier. It is a subjective interpretation, and positivity is considered to be the presence of two or more bands in the CSF, immune-fixing as IgG, and not present in matching serum. Recognized patterns ( Fig. 45.7 ) are normal (banding not detected in either CSF or serum), positive (bands present in CSF but not serum), matching bands present in both serum and CSF (indicative of systemic inflammation), and bands present in both serum and CSF, though with additional bands present in CSF, which may be interpreted as a positive. The absence of CSF oligoclonal banding does not necessarily exclude the diagnosis of multiple sclerosis and its presence does not necessarily confirm the diagnosis, as other demyelinating conditions can also give rise to CSF oligoclonal banding ( Table 45.2 ).

Evaluation for subarachnoid hemorrhage (bilirubin, xanthochromia)

SAH usually occurs secondary to rupture of a cerebral aneurysm and results in arterial bleeding into the subarachnoid space. It is vitally important that the correct diagnosis is made in the first instance, given that patients may present with further hemorrhage and with a potentially fatal outcome. The demonstration of blood on a CT scan will be positive in 98% of patients with SAH presenting within the first 12 hours. Historically, CT scanning has failed to detect approximately 5% of SAH cases, although the advent of newer generation CT scanners has improved sensitivity significantly. ^, Patients with a positive CT scan usually proceed to cerebral angiography, a procedure to confirm the presence and location of an aneurysm so that it can be treated to prevent further hemorrhage. Those patients with a suggestive history of SAH who have a normal CT scan usually proceed to lumbar puncture, although the reliance on lumbar puncture is diminishing as the sensitivity of imaging modalities improves.

After hemorrhage into the CSF, red blood cells (RBCs) undergo lysis and release oxyhemoglobin, which is converted in vivo, through the action of heme oxygenase enzymes in the leptomeninges in a time-dependent manner to form bilirubin and sometimes methemoglobin. Only bilirubin arises solely from in vivo conversion, although CSF bilirubin will also be increased when CSF total protein or serum bilirubin are increased. Oxyhemoglobin may be introduced into the CSF at the time of a “traumatic tap.” Diagnosis of SAH therefore hinges on the detection of bilirubin in CSF, for which spectrophotometry is the recommended approach. Visual inspection for the yellow discoloration (xanthochromia) imparted to CSF is not considered to be reliable, because the sample may appear to be completely clear to the naked eye despite an abnormal spectrophotometric scan. The UK guidelines provide a sound practical framework for how to obtain appropriate samples and undertake and interpret spectrophotometric scans of CSF. It is worth noting that because bilirubin is produced in a time-dependent manner, the absence of detectable bilirubin does not rule out SAH, particularly if CSF is collected within 12 hours of the onset of the bleed.

The UK guidelines recommend that the specimen for spectrophotometry should be the least blood-stained fraction of CSF taken (usually the last and ideally the fourth). The volume must enable laboratory analysis without dilution. The specimen should be protected from light, and it is recommended that the use of pneumatic tube systems to transport the specimen be avoided. A simultaneous blood specimen should be taken for serum bilirubin and total protein measurement. The timing of sampling relative to that of possible hemorrhage should be no less than 12 hours. The specimen designated for spectrophotometry should be centrifuged at 2000 rpm for 5 minutes within 1 hour of collection. The supernatant should be stored in the dark at 4 °C until analysis.

A zero-order spectrophotometric scan is performed on the supernatant between 350 and 600 nm using a cuvette with a 1-cm path length and an initial full-scale deflection of 0.1 absorbance units (AU) and with further scaling, as appropriate.

The following heme pigments should be identified:

• Oxyhemoglobin: Absorbance maximum between 410 and 418 nm.

• Bilirubin: Either a broad peak in the range of 450 to 460 nm or a shoulder adjacent to an oxyhemoglobin peak if present.

• Methemoglobin: The rarest pigment, and if present, it usually manifests as a broader peak than oxyhemoglobin, occurring between 403 and 410 nm.

Net bilirubin absorbance (NBA) is calculated according to Chalmers’ modification to the original method of Chalmers and Kiley, by drawing a predicted baseline, which forms a tangent to the scan between 350 and 400 nm and again between 430 and 530 nm ( Fig. 45.8 ). The absorbance of the scan is measured above this predicted baseline at 476 nm; this is the NBA. If the baseline forms a tangent to the scan before 476 nm, then the measured NBA is by definition zero. The absorbance of any oxyhemoglobin peak above this predicted baseline should also be measured as the net oxyhemoglobin absorbance. Typical zero-order spectra are shown ( Fig. 45.9 ) with an interpretative algorithm ( Fig. 45.10 ) as defined in the UK guidelines. An NBA of greater than 0.007 AU is considered supportive of SAH, although if a visible oxyhemoglobin peak is not observed, an adjustment needs to be made for prevailing plasma bilirubin and CSF protein to derive an adjusted NBA, which is then interpreted according to the algorithm (see Fig. 45.10 ). The predicted absorbance (PA) of CSF at 476 nm because of bilirubin can be calculated according to the following equation and used to derive an adjusted NBA:

Then adjusted NBA = measured NBA-PA.

Most cases that are positive for bilirubin in the first week after an SAH will also be expected to have a visible oxyhemoglobin peak. A traumatic tap introduces red cells into the CSF and results in fluid that is blood stained initially, though it clears as the collection proceeds from tubes 1 to 4, which can be confirmed by measuring cell counts. The distinction may be confounded when a traumatic tap occurs in the presence of an SAH. In a traumatic tap, bilirubin is not produced, given that this is an in vivo process, although a large oxyhemoglobin peak (>0.1 AU) may obscure the region where bilirubin absorbance is normally measured and render interpretation difficult. A series of interpretative comments is suggested in the UK guidelines. Although the wording of the suggested comments in some cases is nuanced, these are essentially dichotomous, channeling a patient toward either angiography or an alternative diagnosis and possibly early discharge.

The UK guidelines recommend caution for use where the CSF bilirubin is increased because of an increased CSF protein alone, or where there is an increased serum bilirubin and the CSF protein is greater than 1.0 g/L, although this may still be consistent with SAH.

Scanning spectrophotometry, however, is not necessarily suited for all clinical laboratories, and many laboratories still rely on visual inspection. As an alternative approach, some have evaluated the measurement of CSF bilirubin on an automated instrument, using the Jendrassik-Grof method, calibrated to measure lower bilirubin concentrations, and found 100% negative predictive value for an abnormal scan at a CSF bilirubin cutoff of 359 nmol/L (0.02 mg/dL) and with no evidence that any case of SAH had been missed. It also provides an alternative approach when there is an insufficient volume of sample for spectrophotometry (a minimum of 500 μL required), although the positive predictive value for an abnormal scan is only 70%. Some laboratories have adopted this approach as a screen to rule out samples that do not need to undergo spectrophotometry. The caveat is that each laboratory should meticulously validate its own diagnostic performance before adopting this procedure into clinical application. ^,

Cerebrospinal fluid biomarkers of dementia

In the setting of dementia, CSF examination has traditionally been used to exclude infection and malignancy, although more recently, CSF analysis using immunochemical techniques has enabled a range of neuronal-specific biomarkers to be measured. CSF β-amyloid (Aβ1-42) is thought to be one of the key biological forms of β-amyloid protein in brain tissue and inversely reflects brain amyloid burden, with reduced concentrations in Alzheimer disease (AD). Reduction of CSF Aβ-1-42 has been shown to occur years before symptom onset and to have positive predictive value for progression from mild cognitive impairment to clinical AD. CSF concentrations of total tau (T-tau), which reflects the intensity of neuronal degeneration and hyperphosphorylated tau (P-tau181P), which is believed to be a marker of neurofibrillary tangle pathologic conditions are both increased in AD. T-tau, however, is also increased after stroke, in inflammatory conditions, and in other neurodegenerative diseases, most notably in Creutzfeldt-Jakob disease, in which concentrations are often orders of magnitude higher than in AD. In contrast, P-tau181P increase is thought to have high specificity for AD. The combination of low CSF Aβ1-42 and increased T-tau and P-tau181P to Aβ1-42 ratio has been used to support the diagnosis of AD, with one study suggesting that the ratio of T-tau to Aβ1-42 ratio is the most robust biomarker combination.

These biomarkers have been incorporated into the revised diagnostic criteria of AD and are also well established as part of inclusion and exclusion criteria for clinical trials. A low CSF Aβ1-42, however, does not always reflect brain amyloid deposition, being seen in other non-AD causes of dementia such as Lewy body disease or vascular dementia. Most importantly, however, it has become recognized that altered concentrations of CSF Aβ1-42 and other markers may result from variations in the way that CSF is collected, stored, or processed, which has hampered the comparison of CSF Aβ1-42, T-tau, and P-tau181P concentrations between different laboratories and studies. The Alzheimer’s Biomarkers Standardization Initiative has made recommendations for preanalytical stringency for CSF biomarkers for AD. Temperature of freezing, delay until freezing, and freeze-thaw cycles have been shown to influence CSF biomarker concentrations, stressing the need for harmonized operating procedures for preanalytical sample handling, although there is no need to recommend any particular time of day for LP. Fasting is not required for analysis of Aβ1-42, T-tau, and P-tau181P biomarker concentrations.

From an analytical perspective, standardization efforts from the IFCC have enabled the production of certified reference materials (CRM) CSF for Aβ1-42, with concentrations assigned using a mass spectrometry reference measurement procedure that has been extensively validated. This has paved the way for the development of fully automated assays with the potential for wider adoption of these biomarkers into clinical practice. Further impetus in this area is likely following the publication by the National Institute on Aging and Alzheimer’s Association (NIA-AA), which has defined AD as a process identified primarily using biomarkers.

There is also an evolving evidence base for autoimmune neurologic disorders manifesting with features of a rapidly progressive dementia, which may be responsive to immunomodulatory therapy and for which the detection of CSF biomarkers of autoimmunity may be helpful. For example, detection of neural autoantibodies in serum or CSF may help inform the physician of a possible autoimmune cause and raise suspicion for a paraneoplastic cause. Available in only a few reference centers, some of the antibodies are predictive of a malignant cause, for example of small-cell lung carcinoma (antineuronal nuclear antibody-type 1 [ANNA-1], collapsin response-mediator protein-5 neuronal [CRMP-5]-IgG), ovarian teratoma ( N -methyl-d-aspartate receptor NMDA-R), and thymoma (CRMP-5 IgG).

Other clinically significant analytes in cerebrospinal fluid