35: Lumbar Puncture (Spinal Tap)

1. Obtaining cerebrospinal fluid (CSF) for the diagnosis of central nervous system (CNS) disorders such as meningitis/encephalitis. Infections that can be diagnosed are bacterial, viral, fungal, and TORCH (toxoplasmosis, other [usually syphilis], rubella, cytomegalovirus, and herpes simplex virus). Meningitis can be present in as many as 15–25% of cases of neonatal sepsis.

2. Aid in the diagnosis of intracranial hemorrhage. CSF studies are indicative but not diagnostic for intracranial hemorrhage: large number of red blood cells (RBCs), xanthochromia, increased protein content, and hypoglycorrhachia (abnormally low CSF glucose content).

3. Diagnose an inborn error of metabolism. CSF amino acid analysis can be obtained to rule out nonketotic hyperglycinemia. Postmortem CSF (1- to 2-mL specimen, frozen) is recommended: suspected inborn error of metabolism.

4. Draining CSF in communicating hydrocephalus associated with intraventricular hemorrhage. (Serial lumbar punctures for this are controversial.) Cochrane review states that early repeated CSF tapping cannot be recommended for neonates at risk of developing posthemorrhagic hydrocephalus.

5. Administration of intrathecal medications. Chemotherapy, antibiotics, or anesthetic agents or contrast material.

6. Monitoring efficacy of antibiotics used to treat CNS infections by examining CSF fluid.

7. Diagnose CNS involvement with leukemia.

8. For the initial sepsis workup (controversial). If CNS involvement is suspected or blood cultures are positive, some recommend a lumbar puncture (LP). Because signs and symptoms of neonatal meningitis are so vague and unspecific, some clinicians advise that all infants with proven or suspected sepsis undergo LP.

Lumbar puncture kit (usually contains three sterile specimen tubes; four sterile tubes are often necessary); sterile drapes; sterile gauze; 20-, 22-, or 24-gauge 1.5-inch spinal needle with stylet (do not use a butterfly needle, as it may introduce skin into the subarachnoid space and form a dermoid cyst); 1% lidocaine; 25- to 27-gauge needle, 1-mL syringe; sterile gloves; mask; gown; hat; and skin disinfectant (10% povidone-iodine solution).

1. Contraindications include increased intracranial pressure (risk of CNS herniation), uncorrected bleeding abnormality, severe bleeding diathesis, infection near puncture site, severe cardiorespiratory instability, and lumbosacral abnormalities that may interfere with identification of key structures.

2. If significant increased intracranial pressure is suspected obtain a computed tomography (CT) or magnetic resonance imaging (MRI) of the head. Herniation rarely occurs in the neonate with open cranial sutures, but is reported.

3. Pain management

   1. AAP recommends that topical anesthetics (EMLA [eutectic mixture of lidocaine and prilocaine] or other topical agents) be applied 30 minutes before the procedure. Nonpharmacologic pain management, if appropriate, can be used.

   2. Lidocaine 0.5–1% (in a 1-mL syringe with a 25- or 27-gauge needle) can be injected subcutaneously. Note: Physiologic instability is not reduced with lidocaine use and is not recommended by some sources.

   3. Systemic therapy. Other recommendations include sedation with a slow IV opiate bolus if the infant is intubated; if not intubated, a bolus of midazolam in a term infant can be used (see Chapter 76).

4. An assistant should restrain the infant in either a sitting or a lateral decubitus position, with the spine flexed as in Figure 35–1, depending on personal preference. An intubated, critically ill infant must be treated in the lateral decubitus position. Some advocate that if CSF cannot be obtained in the lateral decubitus position, the sitting position should be used. In the lateral decubitus position, the spine should be flexed (knee-chest position). The neck should not be flexed because of an increased incidence of airway compromise; maintain airway patency. Supplemental oxygen can prevent hypoxemia. Monitor vital signs and pulse oximetry during the procedure.

5. Once the infant is in position, check for landmarks. (See Figure 35–1.) Palpate the iliac crest and slide your finger down to the L4 vertebral body. Then use the L4-L5 interspace (preferred LP site to avoid cord penetration) as the site of the lumbar puncture. Make a nail imprint at the exact location to mark the site.

6. Prepare the materials. Open sterile containers, pour antiseptic solution into the plastic well located in the lumbar puncture kit.

7. Put gloves on and clean the lumbar area with antiseptic solution, starting at the interspace selected. Prep in a widening circle from that interspace up and over the iliac crest.

8. Drape the area with one towel under the infant and one towel covering everything but the selected interspace. Keep the infant's face exposed. Palpate again to find the selected interspace.

9. Insert the needle in the midline with steady pressure aimed toward the umbilicus.

   1. Guidelines for spinal needle depth. 1–1.5 cm in term infant, <1 cm in a preterm, or calculate depth of needle insertion: 0.03 × body length (in centimeters). Advance the needle slowly and then remove the stylet to check for appearance of fluid. The fluid should be clear but may be slightly xanthochromic (common and associated with maternal labor preceding delivery).

   2. Remove the stylet frequently to keep from going too far and getting a bloody specimen. Early stylet removal improves the success rate. A “pop” as the ligamentum flavum and dura are penetrated is not usually found, as is the case with older children and adults. Rotate the needle if no fluid is seen, and never aspirate with a syringe.

10. Collect 0.5 – 1 mL of CSF in each of the 4 sterile specimen tubes by allowing the fluid to drip into the tubes. CSF white blood cell (WBC) and glucose values can decrease over time (>2 hours); therefore, these need to be sent immediately. For routine CSF examination, send 4 tubes of CSF to the laboratory in the following recommended order:

    1. Tube 1. Gram stain, bacterial culture, and sensitivity testing.

    2. Tube 2. Glucose and protein levels. Others if metabolic disease suspected.

    3. Tube 3. Cell count and differential.

    4. Tube 4. Optional and can be sent for rapid antigen tests for specific pathogens (eg, group B Streptococcus) or PCR (polymerase chain reaction; eg, herpes).

11. For the treatment of communicating hydrocephalus with intraventricular hemorrhage. Remove 10–15 mL/kg of CSF.

12. If a bloody specimen is obtained in the first tube:

    1. Observe for clearing in the second and third tubes. An RBC count on the first and last tubes determines if there is a difference in the number of RBCs/mm³. If the last tube has fewer RBCs than the first, the tap was likely traumatic (usually puncture of the epidural venous plexus on the post surface of the vertebral body). Note: Adjustment of WBC counts in a traumatic lumbar puncture does not aid in the diagnosis of meningitis in neonates.

    2. If blood does not clear but forms clots. A blood vessel has probably been punctured. Because CSF has not been obtained, a repeat tap must be done.

    3. If blood does not clear and does not clot and there are equal numbers of RBCs in the first and last tubes. The infant probably has intracranial bleeding.

13. Replace the stylet before removing the needle to prevent trapping the spinal nerve roots. Withdraw the needle. Maintain temporary pressure and clean and bandage on the site.

14. A repeat tap in 24 to 48 hours is recommended. If the first tap is not diagnostic and the clinical picture is concerning.

There is no evidence that LP-related headache occurs in infants.

1. Contamination of CSF specimen with blood. (See Section III.L, above.) Correction of WBCs for RBCs is not useful. Repeat LP in 12–24 hours.

2. Infection. Sterile technique reduces risk. Bacteremia may result if a blood vessel is punctured after the needle has passed through contaminated CSF. Meningitis can occur if LP is performed during bacteremia. Abscess (spinal and epidural) and vertebral osteomyelitis are rare.

3. Intraspinal epidermoid tumor. Results from performing LP with a needle that does not have a stylet. The cause is the displacement of a “plug” of epithelial tissue into the dura. The incidence of traumatic lumbar puncture is not reduced by the use of a needle without a stylet. Do not use a butterfly or any needle without a stylet.

4. Herniation of cerebral tissue through the foramen magnum. Uncommon in neonates because of the open fontanelle.

5. Spinal cord and nerve damage. To avoid this complication, use the L4-L5 interspace. Between 25 and 40 weeks' gestation the spinal cord terminates between the second and fourth lumbar vertebrae. After 2 months post term, the cord is in the normal adult position.

6. Intramedullary hemorrhage resulting in paraplegia. It is important to consider the location of the conus medullaris in a preterm infant.

7. Bleeding/hematoma. Spinal epidural hematoma, intracranial or spinal subdural hematoma, and intracranial or spinal subarachnoid hematomas have all been reported.

8. Cerebrospinal fluid leakage. Frequent complications are seen on sonograms.

9. Apnea and bradycardia. Sometimes occur from respiratory compromise caused by the infant being held too tightly during the procedure.

10. Hypoxia. Commonly seen; increasing the oxygen during the procedure may help. Preoxygenation may also help.

11. Cardiopulmonary arrest.

Normal CSF values are listed in Table 35–1. Remember that neonatal meningitis can occur with normal CSF values. No single CSF value can exclude CSF meningitis. CSF values in the neonate that indicate meningitis are controversial. Use caution when interpreting results in the premature infant. Data suggests that results in premature infants cannot be reliably used to exclude meningitis.

1. Bloody CSF fluid. See Section III.L.

2. CSF protein and WBC count. Decrease with increasing postnatal age. CSF protein decreases by about 6.8% with each week of age.

3. Elevated CSF protein without increased CSF WBC counts. Seen in congenital infections, intracranial hemorrhage, and para-meningeal infections (eg, brain abscess).

4. CSF WBC counts. Higher in gram-negative meningitis than gram-positive meningitis.

5. Number of bands in a CSF specimen. Does not predict meningitis.

6. CSF glucose. This is 80% of term blood glucose, 75% of preterm blood glucose. A low CSF glucose has the greatest specificity for meningitis.

7. CSF glucose and protein. Cannot accurately diagnose meningitis.

8. Generally CSF protein is higher in preterm infants when compared to term infants. For CSF glucose, values tend to be higher or similar in preterm infants compared to term infants.

9. Other values suggestive of meningitis

   1. CSF WBC count >20–30 cells with predominance of polymorphonuclear (PMN) leukocytes (with bacterial meningitis: infants >34 weeks, median WBC 477/mm³; infants <34 weeks, median WBC 110/mm³).

   2. CSF protein >150 mg/dL in preterm infants, >100 mg/dL in term infants (96% of infants with meningitis have a CSF protein >90 mg/dL).

   3. CSF glucose <20 mg/dL in preterm infants, <30 mg/dL in term infants.

   4. Meningitis can occur with normal CSF values.

10. Values suggesting meningitis is not present

    1. CSF mean WBC count in preterm and term <10 cells/m³.

    2. CSF protein in term infants <100 mg/dL. With preterm infants it varies with gestational age.

    3. Age-specific protein values, mean 0–14 days, 79 mg/dL; 15–28 days, 69 mg/dL; 29–42 days, 58 mg/dL.