In this study the majority of patients (4/5) presenting with symptoms of an impacted fish bone had no demonstrated pathology, and their symptoms resolved in 48 hours. The authors postulate that their symptoms were due to a minor abrasion or other local trauma, but they did not totally discount the fact that a fish bone could have been culprit yet passed spontaneously within the following 48 hours. One of every five patients (20%) did have an impacted fish bone, and the majority of these were easily identified and removed on the initial visit. The most common area for the bone to lodge was in the pharynx, such as the base of the tongue (8/15) or the tonsil (3/15). The authors caution that saliva strands can either obscure or impersonate a fish bone, and they advise caution when attempting to extricate the suspected foreign body under direct vision. The authors believe that the plain x-ray is of limited value in a presence of a normal physical examination. One may see indirect evidence of a fish bone (soft tissue swelling or subcutaneous air), but these radiographic signs may take 12 hours to fully develop.
It's concluded that a fish bone can be found in only one of five patients that present with symptomatology; but if the bone is present, it's often easily visible in the oropharynx. If the bone is not identified on direct examination, a lateral neck x-ray is suggested, but one should not rely on a negative radiograph to rule out a retained bone. If the x-ray and initial physical exam are normal, the authors believe it's safe and reasonable to discharge these patients to return at 48 hours if symptoms persist. They specifically state that referral for immediate endoscopy is not required if only minor symptoms are present.
A Prospective Study of Fish Bone Ingestion: Experience with 358 Patients, Ngan JH, et al, Ann Surg, 1990;11:459
This study from the University of Hong Kong is a prospective evaluation of 358 patients who complained of the following symptoms after eating fish: foreign body sensation, pain on swallowing, rest pain, blood-tinged saliva, or feeling of obstruction. All patients had an initial oral examination and if a bone was not identified, they were evaluated with a plain lateral neck radiograph and fiberoptic endoscopy. All patients were followed until the symptoms completely subsided.
Fish bones were found in 117 of 358 patients (33%). The initial oral exam permitted visualization and removal of the offending bone in 21 cases (6%). Flexible fiberoptic endoscopy was performed in the remaining 281 cases. With this procedure, 94 bones were identified, 82 were retrieved, and 12 bones were dislodged and swallowed during the procedure. Two bones were initially missed by endoscopy. One was removed when persistent symptoms prompted a repeat endoscopy, and this bone was found lodged in the posterior tongue. In the second case, a bone was noted on the radiograph and required repeat endoscopy under general anesthesia for removal. Interestingly, three patients who had ingested fish heads had a large triangular fish bone lodged in the hypopharynx. These were easily seen on x-ray, and all were recovered with rigid endoscopy, but one patient suffered a mucosal tear and two required a lengthy procedure to retrieve the foreign body.
Of the 35 patients who refused initial endoscopy, one developed a retropharyngeal abscess within two weeks. The other 34 patients were asymptomatic at four weeks follow-up.
In this study, the predictive value of symptoms was quite poor. Although the majority of patients with a fish bone had a sensation of a sharp pricking pain on swallowing (predictive value of 76%), the usefulness of any of the other symptoms in actually predicting the presence of a fish bone was questionable, with an overall predictive value of less than 50%. Likewise, the value of a plain radiograph was marginal; it had only a 32% sensitivity. However, when a bone was seen on x-ray, it was 91% specific with an overall predictive value of 66%. Soft tissues and calcified laryngeal structures were misinterpreted as foreign body in 18 of 54 positive radiographs. Therefore, only 31 percent of bones were identified by x-ray.
Interestingly, when a foreign body was present, the patients could localize the area of lodgment by pointing to the general location in the mouth, neck, or chest. On endoscopy, bones were found close to the anatomic vicinity indicated by the patient.
The authors conclude that most swallowed fish bones become lodged in the oral cavity or pharynx, and a good percentage (27%) can be removed directly without formal rigid endoscopy. A complete oral examination is mandatory, and when combined with flexible endoscopy, the vast majority of fish bones can be identified and removed without morbidity. Once fish bones are in the hypopharynx or upper esophagus, they can be associated with the need for lengthy procedures, treatment failures, and mucosal tears during removal. The authors believe that flexible endoscopy is required in all patients complaining of an impacted fish bone if it cannot be visualized on direct oral examination. No specific time frame for such evaluation is advanced.
Comment: All patients who complain of a foreign body in the throat should be taken seriously. Nonchalantly attributing even vague symptoms to anxiety or a benign mucosal scratch can be disastrous. Even relatively smooth or rounded objects that remain impacted in the esophagus have the potential for serious problems, and a fish bone can perforate the esophagus in only a few days. As noted by the authors of the first study, as early as 1936, the famous physician Chevalier Jackson appreciated the potential for serious complications when the physician fails to consider the possibility of a foreign body, does not elicit a proper history, is skeptical of the patient's claim, treats the condition with an apathetic attitude, incorrectly awaits spontaneous passage, or attempts the attribute signs and symptoms to other medical conditions. Although the authors of these two studies have somewhat different opinions about the work-up of a suspected fish bone, both support similar clinical approaches.
Most patients who present after eating fish with a sensation of a bone caught in the throat will not have a bone identified after a complete evaluation. The incidence of actually finding a bone in these two reports was low — 21 percent and 33 percent. Therefore, about three-quarters of the patients who come to the emergency department will not have a bone identified. It is impossible to say how many of these patients actually have a bone that is not documented on the initial visit, but spontaneously passes over the next few hours or days.
Bones may pass, but they do not dissolve while impacted in the throat. The current thinking is that in the absence of a proven retained foreign body, the sensations described are due to minor trauma of the GI tract that are produced when the bone is swallowed. Even though the symptoms are attributed to a mucosal tear or minor laceration, it is extremely uncommon to identify such pathology clinically. Because the GI mucosa regenerates so quickly, it is not uncommon for a superficial injury to heal within 48 hours, so this mucosal trauma theory may be correct.
Some physicians place diagnostic value on the response to swallowing viscus lidocaine or symptomatic relief following the use of other topical anesthetic solutions. If the pain goes away it confirms the suspicion of minor trauma and rules out a foreign body. I have not seen this contention verified in the literature; however, I would caution that this is a dangerous assumption, and just because the pain disappears when someone swallows a local anesthetic, one should never assume that a foreign body is absent. Anesthetics can provide initial relief or facilitate examination, but they have no diagnostic role, and their repeated use is unwarranted and potentially dangerous. A 2% viscous lidocaine solution contains 20 mg per ml of absorbable lidocaine. Because the average adult swallow is 15 ml, each bolus of this anesthetic contains 300 mg of lidocaine. It's well absorbed orally, and can quickly lead to toxicity if the patient is given a bottle to take home.
The ideal ED approach is first to carefully examine the patient's oral cavity. This is best done with the patient sitting in a chair, the use of a bright head lamp, a tongue blade, and spraying the pharynx with an anesthetic. I prefer to use 10% lidocaine spray, but benzocaine is a good alternative. I have the patient gargle for a few seconds, hold it in the throat, and then swallow the anesthetic. Because many fish bones are seen in the oral pharynx with just a flashlight and tongue blade, it's justified to avoid the knee-jerk response of ordering an x-ray before you examine the patient.
I have personally removed fish bones stuck in a tonsil, and the report by Knight and Lesser indicates that the majority of impacted bones can be seen on initial examination. These authors also are very correct in stating that strands of saliva can mimic a fish bone, and therefore one must be very careful about blindly grasping at the posterior pharynx with a hemostat in an effort to pick out an imaginary bone. Interestingly, patients are able to identify the general location of pathology in the cervical area, so look in these areas more carefully. A number of authors have demonstrated that patients can accurately tell if a foreign body is on the right or left side, and can identify the general level of impaction. If the patient points to the sternal notch, you will not see it, but if it is in the submandibular area, a careful direct inspection may be quite rewarding.
The literature is divided in its opinions about the need for immediate endoscopy. Fiberoptic nasopharyngoscopy now allows all patients to be examined in the ED on their initial visit. This is a technique that is becoming more available, and is easily mastered by the emergency physician; I urge its routine use. It's a lot easier to perform than a mirror exam, and indirect mirror laryngoscopy by most emergency physicians is a waste of time. You may see a bone with the NPL scope, but you cannot remove it.
Some physicians will examine the supine patient's oropharynx and larynx with the standard curved blade of the laryngoscope that is used for intubation. This allows a good view of the structures, but not all awake patients can tolerate this examination. It's a reasonable option in the middle of the night if the fiberoptic scope is unavailable. If the initial work-up is normal, I will advise the patient to eat solid food in hopes that it will pick up and dislodge an occasional unseen bone. I have not seen this studied, but patients do it reflexively all the time at home, and it's a common recommendation. I see no harm in doing it.
It's best to work out a prospective arrangement with your ENT consultant, but there is general agreement that if the direct visualization and fiberoptic examination and plain radiographs are negative, the minimally symptomatic patient can be safely discharged. All should have a 24- to 48-hour revisit. Clearly the majority of patients with an initial negative examination will be asymptomatic in two to three days. Although there is always that nagging question of a retained foreign body that becomes asymptomatic only to announce its presence with a septic event or an aortic rupture later, there is general agreement that asymptomatic patients need not be endoscoped. Certainly if the patient is spitting blood, cannot swallow, or has respiratory involvement, immediate endoscopy is indicated. Note that occasionally a repeat endoscopy may be required should symptoms persist. The CT scan has essentially negated the need for soft tissue radiographs.
Radiology of Fish Bone Foreign Bodies in the Neck, Karr AJ, J Laryngology Oralogy, 1987;101:407
This is a brief article that describes two patients with pain in the throat after eating trout or halibut. Both had negative soft tissues x-rays of the neck yet laryngoscopy revealed a fish bone in each case. Based on these cases, the authors experimentally x-rayed the bones of 17 common species of fish in a cadaver model to determine radiopacity. Fish of the cod, haddock, halibut, and monk fish variety had consistent radiopaque bones but mackeral, herring, salmon, and trout demonstrated no radiopacity. It did not matter if the bone was fresh or cooked. This author presents a table showing the relative radiopacity of bones of various species, but this is of little clinical help on any individual patient.
Comment: This article is frequently quoted, but I believe it has significant limitations. Specifically, all of the halibut bones were considered radiopaque when examined in the cadaver, yet one of the patients given as an example had ingested halibut and had a negative x-ray. Physicians tend to overestimate the ability of plain radiographs to detect small fish bones in the throat. Essentially, a lateral neck film is a waste of time on the initial visit for most patients with minimal symptoms. By the time indirect evidence of a bone is seen (air, soft tissue swelling), there is likely a serious complication.
Clearly a negative x-ray does not rule out a retained fish bone. Although Goldman states that about 75 percent of bones will be visible on x-ray (Ann Otol Rhinol Laryngol 1951;60:957), probably the best accuracy one can hope for is about 30 to 40 percent with the type of fish eaten today. One is often frustrated trying to figure out if that little speck of calcium is a calcified ligament, a fish bone, or some other artifact on the x-ray. It helps to have a book of normal radiology variants handy. In the Ngan et al paper, 18 of 54 (33%) radiographs were false positives. Large bones may be obvious on the film, but particularly in elderly patients where laryngeal structure tend to calcify, a lateral neck x-ray has very limited value. One may be lucky enough to see soft tissue swelling or entrapped air — indirect evidence of a retained foreign body — but this may not occur for the first 12 hours. Therefore I don't hesitate to repeat the film if one was negative two days earlier and symptoms persist. There seems to be no particular added benefit of xeroradiography (Radiology 1979;133 218).
If a radiographic study is to be performed, the only one to trust is the CT scan. All other tests (lateral neck, barium swallow) are antiquated and not sensitive enough to be of sufficient clinical value. However, occasionally even the CT scan will be negative, so caution is always advised. Paine et al (Laryngoscope 1999;109:1955) investigated the value of the CT scan to find fish bones in the cricopharyngeus area, a site that often results in perforation. In a cadaver model, 30 bones (size 16x7 cm to 23x7 cm) from various fish were inserted, and the model was subjected to a third generation CT scanner and plain lateral neck x-rays. A blinded radiologist easily identified 29 of 30 bones by CT, and saw the remaining bone with close inspection. With plain x-rays, 19 of 30 bones were not visualized at all, and only four bones were easily seen. A CT scan does not have to be done initially on every patient with a scratchy throat, but this study should convince any physician that the CT scan is the best radiolographic way to find a bone. If you are going to order an imaging study, forget the lateral neck x-ray.
A number of fancy radiology tests have been devised that seem to offer no great benefit, although some physicians continue to use them. The barium swallow may outline a fish bone, but coating the throat with thick layer of barium will make endoscopy very difficult. Others have suggested swallowing a piece of cotton soaked in barium to limit this coating effect with the hope that the ball itself or a few radiopaque strands of cotton will hang up on a foreign body. Importantly, if a barium swallow is to be done with a cotton ball, it should be done under fluoroscopy so any changes of motility of the cotton ball can be evaluated. Some have suggested a barium swallow followed by a glass of water to wash out residual barium hoping that a few flecks will remain on the foreign body. There is a lot of folklore surrounding a barium study or the cotton ball technique, and I have never seen them evaluated prospectively. I personally think it is a myth and a waste of time. Given to the availability of endoscopy and CT scans, I don't use them.
Summary: A number of reasonable options exist, but I suggest the following approach to patients with a history of eating fish who develop subsequent foreign body sensation and pain upon swallowing. First of all, even though less than half have an impacted foreign body, believe the patient when he tells you there is something caught in his throat. Examine the patient under proper lighting, local anesthesia, and positioning. Ask for localization of the sensation, and concentrate your search there with the expectation that you may well see the offending foreign body with only a flashlight and tongue blade. Look carefully in the tonsils and base of the tongue. Most bones end up in the pharynx, not the esophagus.
Don't rely on the relief of pain with a local anesthetic to signify that the symptoms are due to minor mucosa trauma. If initial examination is negative, you can obtain a plain lateral neck x-ray and ask for soft tissue technique, but don't rule out the foreign body if the x-ray is negative. If the symptoms are severe, there is dysphasia, bleeding, or other concerning signs and symptoms, do a CT scan of the neck. It's clearly your best bet to find or rule out a bone. With a minor scratchy sensation, it's appropriate to forgo an initial CT. Master the technique of fiberoptic nasopharyngoscopy (and perform it routinely on the initial visit). If these steps are negative, ED discharge is reasonable if symptoms are minor and follow-up is available in 48 hours. In the meantime, let the patient eat normally.
Aggressively evaluate patients whose symptoms persist for more than 48 hours, even if it means a repeat endoscopy. Remember, retained fish bones do not dissolve. All such patients should be seen on a return visit by your ENT consultant. Bones in the tongue or tonsils are not usually problematic, but bones in the cricopharyngeus area or lower esophagus (very rare) are bound to cause trouble if not removed in a few days.
Fish Bones in the Throat: Suggested Guidelines for the Emergency Department
© 2001 Lippincott Williams & Wilkins, Inc.
- Only about 20 to 30 percent of patients with the sensation of a retained fish bone will actually have one, but you must take this complaint seriously and consider a bone present until proven otherwise.
- Most bones get hung up in the oral/hypopharynx and rarely pass into the distal esophagus. Many are visible on direct vision, especially in the base of the tongue and tonsils.
- No specific sign or symptoms will consistently rule in/rule out a retained bone.
- Patients can generally localize the position of the foreign body.
- A lateral neck x-ray is of little clinical value in most patients with minor symptoms. A CT scan is the test to order if additional investigation is required.
- If the direct exam and fiberoptic exam are normal and the symptoms minor, the patient can be discharged to eat normally and return in 48 hours for a recheck.
- Aggressively pursue symptoms that persist for more than 48 hours.
- Fish bones impacted in the esophagus do not dissolve, they perforate.