xRead - September 2022

Oakley et al.

procedure for the safest and most effective results. A search of the use of IF in the setting of CSF leak confirmation and localization identified 15 relevant studies. Of these, 14 were cross-sectional studies and 1 used mechanism-based reasoning. They reported that IF identified the CSF leak site in 46% to 100% of cases. 7, 46, 70–75 On the other hand, Seth et al. 74 noted that whether or not IF was used intraoper atively did not have a statistically significant effect on the CSF leak recurrence rates. An important consideration in whether or not to use IF is the safety of this diagnostic modality. Studies have indi cated that adverse events are directly associated with the dosage and concentration of IF being administered. Keerl et al. 76 assessed 420 applications of IF ranging from 0.5 to 2 mL of 0.5% to 5% fluorescein, which is equivalent to 2.5 to 100 mg. Moseley et al. 77 surveyed 1111 members of the American Association of Neurological Surgeons, who reported ranges of 0.1 to 5 mL of 5% fluorescein diluted in 0 to 10 mL of the patient’s CSF before intrathecal in jection. Serious adverse events, including lower extremity weakness, seizures, cranial nerve deficits, and even death have been reported to occur with doses as high as 500 to 1250 mg. 76, 77 No intraoperative complications, or minor side effects not definitively attributed to IF alone, are oth erwise reported at doses of 25 to 50 mg. 71, 73, 78, 79 In the United States, the lowest dose and concentration is typi cally used, which is most commonly 0.5 to 2 mL of 0.5% fluorescein (2.5 to 10 mg) diluted in CSF. 79 This literature search also revealed 3 studies that ad dressed the use of topical fluorescein for diagnosis. In this case the fluorescein is placed topically in the nasal cav ity, and when mixed with CSF, changes color to identify the leak. These studies reported an accuracy of 100%. 80–82 Controlled studies using other clear fluids, such as nasal mucus or saline, have not been conducted, so it is unclear if there is the potential for false positives with topical ap plication of fluorescein, such as those seen in the halo test. IF remains a legitimate option for CSF leak confirma tion and localization, but higher-level studies are needed to better outline the safest and most effective uses of this procedure (Supporting material 10). 7, 46, 70–82 Aggregate grade of evidence: C (Level 3: 14 studies; Level 5: 1 study); Benefit: Real-time confirmation and localization; Harm: Invasive study–potential risk to lumbar puncture and intrathecal injection of a neuro-irritative agent; Cost: $524.88; Benefits-harm assessment: Balance of benefit and harm; Value judgments: Low evidence level; Recommendation level: Option; Intervention: Using IF for localization of CSF leaks. Discussion Many modalities exist for the confirmation and localization of a skull-base CSF leak. As this review has shown, they

vary greatly in their benefits, costs, and risks. Although this evidence-based review by its methodology cannot lead to a clinical practice guideline, review of the evidence and the resulting recommendations can be assimilated into a proposed algorithm for the optimal pathway for diagnosing CSF leaks. In 2002, Zapalac et al. 24 reviewed the literature to that point and defined a proposed algorithm for diagnosing CSF leaks. Taking into account accuracy and relative costs, they found the best test to confirm a CSF leak was beta-2 trans ferrin. Once confirmed, the best modality to localize the leak was found to be HRCT. Although a large number of studies have been published since the Zapalac et al. 24 ar ticle, they are almost entirely level 4 evidence; therefore, little has been produced that would lead to questioning these conclusions. Diagnosis of a CSF leak requires 2 steps, confirmation of the leak and localization of the leak. Beta 2 transferrin continues to be the best confirmation test when examining benefits, costs, and risks, at least in the United States. Beta trace protein testing may even be supe rior where available. It is not readily available in the United States. Because of its relatively few benefits and significant costs and risks, this review recommends against RNC’s rou tine use in diagnosis. When beta-2 transferrin testing is not available, MRC is the next best testing modality to confirm a CSF leak. For the second step of localizing the leak, HRCT con tinues to be the study of choice. It is readily available, noninvasive, and highly accurate. Because of its nearly 3-fold higher costs relative to HRCT with similar accuracy, we agree with the Zapalac et al. 24 algorithm and continue to place MRC after HRCT as a localizing study. MRC without contrast may be as efficacious in confirm ing a CSF leak as RNC with similar costs. When intrathecal contrast is used, MRC appears to be superior to RNC in confirming a leak. Because MRC can also localize where RNC cannot, it appears that MRC should replace RNC in confirming the presence of a CSF leak when beta-2 trans ferrin cannot be performed (eg, because of an inadequate sample) or is equivocal. This pathway differs from the Zapalac et al. 24 algorithm published in 2002. All evidence published since 2002 regarding CTC con tinue to support the Zapalac et al. 24 group’s exclusion of this modality from working up a CSF leak. The over all evidence leads to a continued recommendation against use of this study. Because of the significant risks associ ated with IF, the evidence continues to support the use of this modality as an option. Zapalac et al. 24 put it third in line behind HRCT and MRC for localizing and the evi dence since that time continues to confirm IF’s place in the algorithm. In summary, little has been published since 2002 to al ter the Zapalac et al. 24 algorithm. The only change that evidence from the last 13 years mandates is the removal of RNC and its replacement with MRC as a confirm ing study. The resulting updated algorithm is shown in Figure 1.

International Forum of Allergy & Rhinology, Vol. 6, No. 1, January 2016

14