Disclaimer: The information contained within the Grand Rounds Archive is intended for use by doctors and other health care professionals. These documents were prepared by resident physicians for presentation and discussion at a conference held at Baylor College of Medicine in Houston, Texas. No guarantees are made with respect to accuracy or timeliness of this material. This material should not be used as a basis for treatment decisions, and is not a substitute for professional consultation and/or peer-reviewed medical literature. Fish Bone Impactions Fish bones are the most common upper aerodigestive and esophageal foreign body found in adults. However, they can be difficult to diagnose since fish bones are translucent on physical exam and often radiolucent. Complications are fairly rare. Usually these bones pass, but when the complications do happen they can be catastrophic, including neck abscesses, mediastinitis and esophageal aortic or esophagocarotid fistulas. The anatomy of the fish is fairly simple: they have head, a body and a tail. They have fins, which are supported by bony rays. That's why they are known as ray-finned fish. The bones of real interest are the pin bones, which lie between these ray bones and often get stuck in the prepared filets. They are also very small and very sharp and are the bones that most often cause a problem. As I mentioned before, fish bones are the most common adult esophageal foreign body, comprising 60% of cases in a series on Hong Kong. Older children are also at risk once they start eating fish. Chevalier Jackson first mentioned the problem of bone ingestion in terms of chicken and fish bones. He stated that the most common sites of impaction were the tonsil, the posterior tongue, the vallecula, and the cervical esophagus. In the Hong Kong study, out of a total of 117 bones, 73 were in the oropharynx, 18 in the larynx, 18 in the hypo-pharynx and 2 were found in the esophagus. Patients complain mostly of a foreign body sensation. Although dysphagia is certainly not specific to a fish bone, a sharp pricking sensation is highly predictive. When there is a bone present, the side of symptom is predictive of the side of the impaction. Physical exam will certainly catch the majority of fish bones in the oral cavity. The bones can be found in the tonsil and the posterior tongue. Fiberoptic laryngoscopy or mirror exam can be used to locate bones in the vallecula and pyriform sinuses, where they can be fairly easily retrieved with special instruments. However, bones in the cervical esophagus are more problematic, and, in that case radiologic studies are particularly useful. The gold standard of diagnosis is rigid or flexible esophagoscopy. It allows visualization and retrieval. However, it does require general anesthesia and there is a 0.25 to 0.6 risk of esophageal perforation. To prevent any unnecessary esophagoscopies, it is a good idea to order radiologic studies. Bones by themselves are clearly visible on plain film. However, plain films may exhibit poor sensitivity when the bone is lodged in the area of maximum soft tissue overlap. There is also poor specificity because of thyroid cricoid and hyoid calcifications, which can be misleading. . There are a number of studies on plain film diagnosis of fish bone impactions. Dr. Ell found that most bones from fish eaten in Britain were visible. In his most recent study, bones from 14 different fish species were placed in a pig's head and plain films were taken. He found that Herring, Pike, Mackerel and Trout bones were not very visible. Researchers from Ireland placed bones of commonly eaten fish into a human cadaver head and neck. He found that all the bones were visible except for one Brill bone, which was missed in the oropharynx but easily seen in the valleculum. However, in Hong Kong researchers performed a double blind trial of 100 films with proven fish bone impactions and put them together with 100 normal films. After review by a blinded radiologist, plain films had a sensitivity of 25 % and a specificity of 86%, which is fairly poor. The fish most commonly eaten in Hong Kong include the following: Grass carp, Goldfish, Grouper and Golden Thread. None of these fish had been included in either the British or the Irish study. Sundgren from Sweden performed a retrospective study of 42 consecutive patients who presented with suspected fish bone impaction. Only 7 were eventually treated by esophagoscopy and only 2 out of the 7 were diagnosed before going to the operating room, either by plain film or barium swallow. Afterwards, on a retrospective review only 5 out of 7 were seen. The most commonly eaten fish in Sweden are Salmon, Cod, boiled Ling and Herring. So, there are some conflicting reports about the usefulness of plain films, which perhaps varies by region and the species of commonly eaten fish. Another popular study that has been used is a barium swallow with cotton pledgets. In the one series that I found that addressed this test, fifty-three patients in Spain were studied. There was 70% sensitivity and 81% specificity, which is adequate. A major problem with performing a barium swallow is that it coats the esophagus and makes subsequent esophagoscopy or exam very difficult. Recently CT has been reported as being very useful. The first study published on using CT to locate chicken and fish bones was done in Jerusalem in 1993. In a series of 13 patients, they reported that CT clearly demonstrates small bones. They also mentioned that there was some soft tissue injury and inflammation that could be diagnosed through this exam as well. Similarly in Japan, a series of patients had both plain film and CT done. Plain films missed about 56% of Japanese fish bone impactions whereas CT demonstrated the bones in all cases: 11 out of 11. They also made a pig intestine model, putting bones into a pig intestine and performing a CT scan. They were able to visualize all the bones. An Australian group last month took bones from 10 different fish and placed them in the cricopharyngeus of the head and neck cadaver and took both plain films and CT scans. They had an unblinded radiologist read the films. The radiologist felt that 19 out of 30 bones were not seen by plain films. But 29 out of 30 bones were easily seen by CT. However, there were some shortcomings to this study. First the films were reviewed by an unblinded radiologist. There were no control films used, and they made no attempt to try and correlate the visibility of the bone with the species of the fish. Fish bones are sharp objects and they can get lodged in the aerodigestive tract and cause complications, although this is rare - about 1% to 3% percent out of the series from Hong Kong. But the associated complications are potentially catastrophic, including cervical abscess, mediastinitis, esophagocarotid fistula, lung abscess if it gets into the airway and perforated bowel as it passes through the intestinal tract. Given this information, there are still a few questions. Are the species of fish common eaten in the United States visible on plain films? Do variables such as the particular species of fish or the location and orientation of the bone affect visibility on plain film? How does CT directly compare to plain film? We hypothesized that plain films exhibit poor sensitivity, and specificity for Gulf Coast fish. Low optical density of the fish bone was measured by transmission densitometry. Location in the pyriform sinuses verses the vallecula, and orientation orthogonal to the film would diminish visiblity. Our last hypothesis was that CT would identify all bones regardless of position or orientation. We obtained bones from ten different fish species commonly eaten in the Gulf Coast. We placed a sample of each bone in a head and neck cadaver specimen with a laryngoscope in the vallecula and in the region of the pyriform sinus. We took plain films and CT scans of this experimental setup, and then presented both control films and positive films to blinded readers to calculate sensitivity and specificity. The ten specifies of fish included were: Large Mouth Bass, Catfish, Drum, Flounder, Red Snapper, Redfish, Salmon, Talopias, Tilefish, and Trout. We collected five different bones from at least two different fish from each species. We tried to collect bones that were between 2-cm and 2.5-cm long, since that was the average length of recovered bones from the Swedish study. First, we made a plain film radiograph of just the bones. All the bones were fairly visible on plain film. We used an optical densitometer, which is an instrument that measures the light transmission through a film, and were able to get a quantitative number as to the radiopacity of each bone. We took the average value of each bone as the average optical density of the particular species. Bass have the lowest optical density, meaning that it was the most radiopaque, followed by Catfish, and Red Snapper. Tilefish was the least radiopaque. The real test was to see if the bones were visible in a soft tissue model. We placed a sample of each bone in a human cadaver specimen, in three locations. We placed bones in the vallecula, one parallel to the film and the other orthogonal to the film. We placed another bone in the piriform sinus. Since there were ten different species with three locations, there were a total of thirty positive films. In order to test specificity, we also put in 20 negative films. The films were read by one attending in our department and two residents, and they did about as well as the previously published studies, with a sensitivity of 39% and a specificity of 72%. These numbers have a 10% confidence interval. Over-reading tended to happen in the hypopharynx and cervical esophagus overlying the thyroid cartilage. The cadaver was of a gentleman who was 80 years old and he had extensive calcifications throughout his neck. When we tried to compare the visibility of the different species, we found that the Redfish and Bass were the least visible bones, while Trout and Flounder were highly visible. In fact all the Flounder bones were identified. This histogram compares the optical density with visibility. You can see there is not really a very good correlation between the two bars. As a matter of fact, the correlation coefficient that we calculated was -0 .33. We also found that position did affect the visibility of the bone. Bones in the vallecula were identified 53% of the time as compared to bones in the pyriform, which were seen 27% of the time. As far as orientation, it was surprising to us that bones placed orthogonally to the film were far more visible than bones placed in the same plane. I think this is because the bones placed orthogonally were shot on end, and were therefore more visible. We placed the same experimental subject in the CT scanner, using 3-mm axial slices with a soft tissue protocol. A Neuroradiologist at Ben Taub read these, along with 5 negative CT scans. We found that all the bones were very easily seen, with the exception of one bone that had been placed in the vallecula in the axial plane. This might suggest that thinner slices need to be taken if you are fishing for a bone by CT scan. We also placed a bone submucosally. We found that the Salmon bone placed submucosally in the area of the pyriform sinus was also very easily visualized. In conclusion, plain films exhibit very poor sensitivity and specificity in identifying U.S. fish bone foreign bodies, which was consistent with our hypothesis. However, the optical density of species of fish did not correlate at all with visibility. Location in the pyriform sinus and orientation parallel to the film also diminished visibility. However, CT scanning is a very powerful technique to locate suspected fish bone impactions. Here is a treatment algorithm from a previous study. If physical examination does not find the bone, then plain film of the neck and chest is recommended. If that is negative, and the patient continues to be symptomatic, CT scan without contrast is recommended. If that is negative, the patient can be discharged home without any further worries. The cost of this algorithm is high, though. A reasonable alternative would be to CT scan patients before performing esophagoscopy. A negative CT scan can save a patient a trip to the OR. In summary, fish bone impactions are very common complaints. Missed bones allow rare but potentially catastrophic complications, which is why it is important to locate the bone. Plain films have very poor sensitivity and specificity. CT scans are a powerful diagnostic tool and help may avoid a trip to the operating room. Case Presentation The patient is an otherwise health 35-year-old HF who presents to the emergency room with odynophagia after eating catfish that was prepared at home the previous evening. She denies difficulty breathing or voice changes, and she is able to take solids and liquids. On physical exam, she appears well nourished and in no apparent distress. She is afebrile. Ears: clear. Nose: clear. OC/OP: no evidence of foreign body. Larynx (fiberoptic): no foreign body visible from vallecula to pyriform sinus, no inflammation. Neck: tender to palpation over R thyroid cartilage. A lateral neck film was obtained which showed no evidence of foreign body or soft tissue inflammation. The patient was discharged home from emergency room with follow up the next morning. She did not keep her scheduled appointment. Braverman I, Gomori JM, Polv O, Saah D. The role of CT imaging in the evaluation of cervical esophageal foreign bodies. J Otolaryngol 1993;22:311-314. Dreyer CJ. Demineralization of bone. Nature 1965;207:94. Eliashar R, Dano I, Braverman I, Dangoor E, Sichel J. Computed tomography diagnosis of esophageal bone impaction: a prospective study. Ann Otol Rhinol Laryngol 1999;108:708-710. Ell SR. Radio-opacity of fish bones. J Laryngol Otol 1989;103:1224-1226. Ell SR, Sprigg A. The radio-opacity of fish bones-species variation. Clin Radiol 1991;44:104-107. Ell SR, Sprigg A, Parker AJ. A multi-observer study examining the radiographic visibility of fishbone foreign bodies. J R Soc Med 1996;89:31-34. Evans RM, Ahuja A, Rhys Williams S, Van Hasselt CA. The lateral neck radiograph in suspected impacted fishbones-Does it have a role? Clin Radiol 1992;46:121-123. Herranz-Gonzalez J, Martinez-Vidal J, Garcia-Sarandeses A, Vazquez-Barro C. Esophageal foreign bodies in adults. Otolaryngol Head Neck Surg 1991;105:649-653. Hone SW, Fenton J, Clarke E, Hamilton S, McShane D. The radio-opacity of fish bones: a cadaveric study. Clin Otolaryngol 1995;20:234-235. Jackson CL. Foreign bodies in the air and food passages. Postgrad Med 1948;4:281-290. Nandi P, Ong GB. Foreign bodies in the oesophagus: review of 2394 cases. Br J Surg 1978;65:5-9. Ngan JH, Fok PJ, Lai EC, Branicki FJ, Wong J. A prospective study on fish bone ingestion: experience of 358 patients. Ann Surg 1990;211:459-462. Palme CE, Lowinger D, Petersen AJ. Fish bones at the cricopharyngeus: a comparison of plain-film radiology and computed tomography. Laryngoscope 1999;109:1955-1958. Rombauer, Irma et al. The New Joy of Cooking. New York: Scribner; 1998. Sundgren PC, Burnett A, Maly PV. Value of radiography in the management of possible fish bone ingestion. Ann Otol Rhinol Laryngol 1994;103:628-631. Watanabe K, Kikuchi T, Katori Y, Fujiwara H, Sugita R, Takasaka T, et al. The usefulness of computed tomography in the diagnosis of impacted fish bones in the oesophagus. J Laryngol Otol 1998;112:360-364. Yang CY. The management of ingested foreign bodies in the upper digestive tract: a retrospective study of 46 cases. Sing Med J 1991;32:312-315. Grand Rounds Archive | Department Home page BCM Public | BCM Intranet | Privacy Notices | Contact BCM | BCM Site Map | ©2001-2006 Baylor College of Medicine
|