Critical analysis of the laryngopharyngeal reflux literature

Abstract

We read with interest the article by Kotby et al. [1] which highlights some of the weaknesses of the laryngopharyngeal reXux (LPR) literature. I am referring to their comment on posterior commissure and in particular that “there is no posterior commissure”. The existence of the posterior commissure of the human larynx has been disputed since the 1980s as shown in the work of Hirano et al. [2]. More recently though, Tucker et al. [3] in their embryo-anatomical study concluded that “the posterior end of the glottis is a wall, the cricoid cartilage; the posterior median cricoid cartilage is also the developmental anatomical site of union of the cricoid cartilage and the posterior commissure. It is a relevant, deWnitive, anatomical site, as is the anterior commissure and does not vary with vocal fold mobility”. We agree with the rest of the comments, especially on empiric antacid trials. “Empiric”—from the greek word “empeiriki”—means “from experience”. The evaluation of any medical outcome that relies on empiric treatment and not on accurate diagnostic methods is bound to be Xawed and is—if we want to adhere to evidence based medicine (EBM) terminology—one of the lowest forms of evidence. For years gastroenterologists were talking about the silent manifestations of gastroesophageal reXux (GORD) and in the last 15 years the laryngologists have literally “taken over” the literature writing about LPR and its implication in head and neck symptoms that cannot be otherwise explained. Pepsin has been found in the middle ear and in the larynx and even sinusitis has been linked to reXux [4]. Some classic animal studies have proven with robust methodology the link between acid injury and laryngeal pathology [5]. Nevertheless, physicians have spent years trying to Wnd an ideal test for diagnosing “upper neck” reXux and the norm for interpretation of pH in the proximal oesophagus has never been universally agreed upon. The acid throat family has always included multifactorial symptoms and anything that responds to antacids may equally respond to other treatments. A typical example of empiric treatment based on poor evidence and a well-established worldwide practice over the last 30 years was the attempts to treat globus pharyngeus patients—a classic component of many well-accepted and widely used reXux indexes—with expensive and potent antacids [6]. Globus was never proven to be an “acid phenomenon” [7] and yet the vast majority of family physicians and specialists would often prescribe antacids. If one looks at the literature, there are numerous randomised trials and even more uncontrolled studies on antacids for LPR. Level I evidence though is lacking and those few positive studies are based either on poor methods or on diagnostic criteria that use empiric treatment and symptom/sign questionnaires as main methods for LPR diagnosis. One cannot aVord to be critical of course, as dual probe prolonged pH metry—once thought a gold standard test—is far from a ideal diagnostic tool. In the last few years and following two systematic reviews [8, 9] and a Cochrane metaanalysis [10] of randomised controlled trials (RCTs), it was evident that something is not “right” in the empiric treatment of reXux. Is it the dose, the length of treatment, the sometimes non-acidic nature of reXux or perhaps the timing of PPI dosage? Are we even using the right drugs? No one knows for sure. Comparing placebo to a drug is strong EBM only if you have an equally strong, highly speciWc and sensitive diagnostic tool for the condition you are testing. And at the moment we not only lack this, but also many of the laryngologists do not agree on how to interpret a laryngeal picture of “reXux laryngitis”; even experienced laryngologists often score diVerently “an angry looking larynx”. P. D. Karkos (&) Royal Liverpool University Hospital, 36 Hopkinsons Court, Walls Avenue, Chester CH1 4LN, UK e-mail: pkarkos@aol.com