Pathophysiology and clinical evaluation of iatrogenic cerebrospinal fluid rhinorrhea after sinus surgery

Introduction

Despite significant advances in endoscopic surgical training, technique, and instrumentation, endoscopic sinus and skull base surgery can result in inadvertent skull base injury with subsequent cerebrospinal fluid (CSF) leak a small percentage of the time. Iatrogenic CSF leaks are the most common cause of CSF leak overall in the literature, representing a more frequent cause of CSF rhinorrhea than accidental traumatic injuries. Although traditional endoscopic sinus surgery (ESS) is a safe procedure with a low rate of major complications, ESS-related CSF leaks still account for approximately 25% to 41% of all CSF leaks. ^, While rates of CSF leak per surgery are overall low, a study of otolaryngologists performing ESS showed that 25% of these physicians had experienced an intraoperative iatrogenic CSF leak within the past 5 years. The frequency with which ESS is performed predicates the need for appropriate recommendations and guidance for the prevention of CSF leaks, as well as an evidence-based approach to the evaluation and workup of iatrogenic CSF leak.

Preoperative anatomic risk factors

Specific areas of the skull base are highly prone to injury during ESS, and anatomic variations in these areas can also further contribute to CSF leak risk. CSF leaks most commonly occur during the ethmoidectomy portion of functional ESS, either at the lateral lamella of the cribriform plate or the posterior fovea ethmoidalis. The lateral lamellae can take a variety of angles in the coronal plane off the cribriform plate, and they attach superiorly to the orbital plates via the fovea ethmoidalis, thus making up the roof of the ethmoid air cells that are encountered during ESS. The most common area of injury at the lateral lamella occurs adjacent to the anterior ethmoid artery and posterior to the frontal recess, where the skull base is most fragile, reaching a thickness of 0.05mm or less. Additional sites of potential injury and resultant CSF leak during ESS are the frontal recess as well as the sphenoid sinus.

Given these anatomic regions of interest, it is critical to carefully review all patient imaging before sinus surgery to identify common landmarks and anatomic abnormalities and to evaluate potential areas which will require increased vigilance intraoperatively. The skull base generally slopes inferiorly along the ethmoid roof from the frontal recess to the planum sphenoidale, and the thickness of the anterior skull base or ethmoid roof should be assessed on coronal and sagittal computed tomography (CT) imaging because this will be the least resistive area during dissection. The skull base should also be carefully examined for any areas of soft tissue that might suggest a preexisting dehiscence or encephalocele. Although surgeons often have their own systematic methods for reviewing preoperative imaging, several classification schemes have been developed to further assist with this analysis. Traditionally, the Keros classification has been used to describe cribriform and skull base configurations with increasing depth, with the highest depth classification (class 3) signifying an increased risk for skull base violation with subsequent CSF leak. Keros type 1 configurations correspond to a vertical height of 1 to 3 mm of the lateral lamella, type 2 is 4 to 7 mm, and type 3 is 8 to 16 mm. Importantly, patients may have an asymmetric Keros configuration from one side to the other.

More recently, several groups have highlighted the limitations of the Keros classification and have shown in several case-control studies of iatrogenic CSF leaks during functional ESS that the angulation of the lateral lamella is more predictive of CSF leak risk. Preti et al. proposed a new classification scheme based on this angulation, studying the angle formed between the lateral lamella and the continuation of a horizontal plane passing through the cribriform plate. A steeper slope of the lamella prognosticates potential risks when medial ethmoid air cells are dissected. In this classification, an angle of greater than 80 degrees is considered low risk, 45 to 80 degrees is considered medium risk, and less than 45 degrees is considered high risk. The vast majority of all CSF leaks in this case series occurred in high-risk individuals, and even within subjects who experienced iatrogenic CSF leaks, the side that the CSF leak occurred on had a statistically narrower angle than the contralateral side. These findings corroborate those found by Heaton et al., who in addition to showing that a more pronounced lamella slope predisposes to skull base injury, showed that a low skull base height relative to the hard palate and a steep skull base angle are relevant predictors of iatrogenic CSF leak.

The use of image guidance during ESS has become more ubiquitous in recent years, and many surgeons believe that image guidance is helpful intraoperatively and improves surgeon confidence in anatomic localization. Image guidance can be particularly helpful in cases of chronic rhinosinusitis with nasal polyp disease or during revision surgery. In these cases, anatomic landmarks may be absent, as in the case of prior surgery, or obscured by polyp disease. Extensive polyp disease can also cause thinning and dehiscence of the skull base, further increasing the risk of iatrogenic CSF leak. Image guidance also is useful for surgical safety in individuals with an asymmetric skull base; fewer than 50% of individuals have a symmetric fovea ethmoidalis, and findings on one surgical side may not translate to the other.