Pill-Induced Esophageal Injury: Endoscopic Features and Clinical Outcomes

• Introduction
• Patients and Methods
• Results
• Discussion
• References

Background and Study Aims: Pill-induced esophageal injury is a common but under-reported problem. The purpose of this study was to explore the clinical and endoscopic features, and the outcome of pill-related esophageal injury.
Patients and Methods: Endoscopy records for the period from January 1997 to June 2003 were searched for reports of esophageal pathology. The records of patients with pill-induced esophageal injury were evaluated.
Results: A total of 92 patients with pill-induced esophageal injury were identified (33 men, 59 women; mean age 59, range 25-87). Common symptoms were odynophagia (n = 69, 75 %), chest pain (n = 55, 60 %), vomiting (n = 53, 58 %), dysphagia (n = 31, 33 %), and hematemesis (n = 14, 15 %). The endoscopic findings in the esophagus were: erythema in 76 patients (83 %), erosions in 53 patients (58 %), ulcers in 24 patients (26 %), seven of which were ”kissing” ulcers, esophageal ulcer with bleeding in 17 patients (18 %), and esophageal strictures in seven patients (8 %). The causative pills were nonsteroidal anti-inflammatory drugs in 38 patients (41 %), tetracyclines in 20 patients (22 %), potassium chloride tablets in nine patients (10 %), alendronate in eight patients (9 %), and other drugs in 17 patients (18 %). Underlying diseases included diabetes in 60 patients (65 %), ischemic heart disease in 39 patients (42 %), and hypothyroidism in four patients (4 %). The mean hospital stay was 1.94 days; 14 patients (15 %) required injection of epinephrine 1 : 10 000 to control bleeding; and two patients died.
Conclusions: Pill-induced injury may present as erosions, kissing ulcers, and multiple small areas of ulceration with bleeding, mainly in the middle third of the esophagus. Advanced age, female gender, diabetes, and ischemic heart disease were common associations. The majority of patients made an uneventful recovery.

Introduction

Pill-induced esophageal injury was first described in 1970 [1] [2]. More than a hundred different medicines have been reported to cause esophageal injury. Most patients suffer only self-limiting pain but serious complications can occur. Pill-induced esophageal injuries occur when a caustic medicinal pill dissolves in the esophagus and releases its noxious contents, particularly if transit is delayed. Injury of this type, called ”pill esophagitis” or ”pill-induced esophageal injury”, is common but unfortunately is under-reported. The aim of this study was to evaluate the endoscopic features and clinical outcomes of pill-induced esophageal injuries.

Patients and Methods

We reviewed the records of upper gastrointestinal endoscopies performed between January 1997 and June 2003. The records of patients who were coded for esophageal pathologies (ICD-9-CM) were further evaluated to identify those who had had pill-induced esophageal injury. These patients had presented in outpatient clinics or as an emergency with acute esophageal symptoms. We included data from patients who had a clear history of taking pills and who had experienced an acute onset of esophageal symptoms (heartburn, chest pain, dysphagia, or odynophagia) of less than 3 days' duration, based on the definition of pill esophagitis described by Kikendall [1]. We recorded the demographic characteristics, the pill responsible for the symptoms, endoscopic features, associated systemic illnesses (such as diabetes or ischemic heart disease), and the clinical outcomes (including duration of hospital stay and follow-up findings) of the patients who satisfied the inclusion criteria of our study. Patients with esophageal varices or malignancy were excluded from the study, as were immunocompromised patients, patients who had ingested corrosives, and patients with a long history of gastroesophageal reflux symptoms. Patients with viral or fungal lesions in the esophagus were excluded by history and by esophageal biopsy where this had been found necessary.

Results

A total of 14 223 upper gastrointestinal endoscopies were performed during the 5.5-year study period. Out of these, 1533 reported a finding of esophageal pathology and 92 patients (6 % of those with esophageal pathology) had pill-induced esophageal injury and fulfilled the inclusion criteria (33 men (36 %), 59 women (64 %); mean age 59, range 25 - 87). Common symptoms were: odynophagia in 69 patients (72 %), chest pain in 55 patients (59 %), vomiting in 53 patients (58 %), dysphagia in 31 patients (33 %), and hematemesis in 14 patients (15 %).

Endoscopic findings included: acute superficial erythema (n = 76, 83 %; see Figure [1]), esophageal erosions (n = 53, 58 %; see Figure [2]), esophageal ulcers with exudates (n = 24, 27 %; see Figure [3]), esophageal ulceration with bleeding (n = 17, 19 %; see Figure [4]), and strictures (n = 7, 9 %). Other endoscopic findings included ”kissing” ulcers (ulcers facing each other) in the mid-esophagus in seven patients (8 %; see Figure [5]) and an area of denudation of the proximal esophageal mucosa in three patients (3 %; see Figure [6]). As shown in Table [1], most of these lesions involved the middle third of the esophagus.

Of our 92 patients, 38 had taken a nonsteroidal anti-inflammatory drug (NSAID): 15 patients had taken aspirin; 23 patients had taken other NSAIDs (naproxen (n = 14), indometacin (n = 7), ibuprofen (n = 3). Among the patients with NSAID-induced esophageal injury, 14/38 patients presented with a bleeding esophageal ulcer and 12 of these 14 patients required injection of epinephrine 1 : 10 000 to achieve hemostasis.

Antimicrobial drugs were responsible for esophageal injury in 27 patients in the present series: the most commonly implicated were from the tetracycline group (doxycycline in nine patients, minocycline in five patients, and oxytetracycline in six patients). Other antibiotics which caused esophageal injury were ampicillin, clarithromycin, and rifampicin (see Table [1]). Antimicrobial drugs caused esophageal inflammation in 21 patients and ulceration without bleeding in three patients.

In the present series of 92 patients, underlying systemic diseases included: diabetes mellitus in 60 patients (65 %), ischemic heart disease in 39 patients (42 %), previous cardiac surgery in 11 patients (12 %), and hypothyroidism in four patients (4 %).

The majority of patients were treated conservatively with proton pump inhibitors and/or sucralfate. The offending medicine was stopped in all patients. Blood transfusion was required in the 17 patients whose pill-induced esophageal injury was associated with blood loss. Injection with epinephrine 1 : 10 000 was required in 14 patients (15 %) to stop bleeding. The mean duration of hospital stay was 1.94 days (range 1 - 6 days). Two patients who had taken aspirin died as a result of torrential bleeding and aspiration pneumonia.

All the patients who were discharged from the ward were followed up in the clinic within 2 weeks: their symptoms abated in due course and uneventfully. After a mean follow-up period of 16 ± 6 weeks, the patients showed no evidence of recurrence or persistence of symptoms. Follow-up esophagogastroduodenoscopy was performed in patients who had an esophageal ulcer with bleeding or persistent symptoms. Follow-up esophagogastroduodenoscopy was performed after 2 weeks in all 15 surviving patients out of the 17 patients who had esophageal ulceration with bleeding and in all these patients the esophageal lesions were shown to have recovered completely.

Discussion

This study represents the largest single-center experience of pill-induced esophageal injury reported to date. We have described the endoscopic appearane of various esophageal lesions that were caused by a number of different pills.

This experience of pill-induced esophageal injury was based on patients who presented to a tertiary care center, either in the outpatient clinic or in the emergency room, with significant acute-onset esophageal symptoms of less than 3 days' duration that warranted endoscopic examination. We suspect that there were several other patients with minor complaints or a history of symptoms of more than 3 days' duration who were not enrolled, however. The present series did not include patients who may have developed pill-induced esophageal injury symptoms but who remained unnoticed because their symptoms were less intense. We may therefore be greatly underestimating the scale of this problem, our patients representing only the tip of the iceberg. The mean duration of hospital stay for all our patients, regardless of type of esophageal injury, was 1.94 days, which suggests that most of these patients were emergency patients and representative of a small number of patients who have severe esophageal injury. We have identified 92 patients who have had significant pill-induced esophageal injury in last 5.5 years in our institute. The entire world literature has documented nearly 1000 cases over the past 30 years, an impressive number, but still quite small compared with the estimated case frequency [1]. In Sweden, an annual incidence of four cases of pill esophagitis per 100 000 population was recorded over a 4-year period in the 1970 s [3]. The incidence today is probably much higher because medicines are increasingly prescibed in pill form and as a result of the use of newer medicines such as bisphosphonates.

Oral medications are usually prescribed for adults as tablets or capsules, which has the obvious advantages of ease of ingestion and transportability, compared with the alternative liquid preparations. One of the disadvantages of medications being in tablet or capsule form, however, is the high concentration of active medicine in the pill, which is more likely to injure susceptible tissues, particularly the esophagus. This is particularly true if the pill is ingested without liberal quantities of water or if it is taken in the supine position [4]. Injury of this type is unfortunately common and under-reported. Larger size pills such as clarithromycin, alendronate, and ibuprofen are more commonly retained in the esophagus than smaller size pills. Pills with a sticky surface (some gelatin capsules, such as that covering doxycycline for example) are more commonly retained than pills with a less adherent surface [5].

The precise mechanism of injury for many pills is only speculative. Generally, the contents of many pills are sufficiently caustic to injure the esophagus if retained and released in situ. This has been demonstrated by placing the pills or pill contents directly onto human buccal mucosa or onto the esophageal mucosa in a number of animal models [5]. The drugs which caused pill-induced esophageal injury by this mechanism in our study were potassium chloride and quinidine. Tetracycline, ascorbic acid, and ferrous sulfate produce an acidic solution (pH < 3.0) when dissolved in 10 ml of water, suggesting that they may cause acid burns [3]. Other postulated mechanisms of pill-induced injury include the induction of gastroesophageal reflux by theophylline and anticholinergics, the production of locally high osmolarity by potassium chloride, and intracellular poisoning after mucosal uptake of doxycycline and NSAIDs [6].

In our series, the age range of the patients who suffered pill-induced esophageal injury was 25 - 87 years. Women sustained this type of esophageal injury twice as frequently as men, probably because women in this part of the world are more likely to be treated with potentially injurious agents, such as alendronate for osteoporosis, NSAIDs for rheumatoid arthritis, and antibiotics (notably tetracycline) for acne vulgaris.

Upper gastrointestinal endoscopy is needed to assess the degree of damage caused by a pill, especially in the presence of warning symptoms (such as odynophagia, anemia, anorexia, or hematemesis). Apart from the history of consumption of the offending pills and the acute development of moderate to severe esophageal symptoms, there are few typical features which are diagnostic of pill-induced esophageal injury. Endoscopy usually shows focal areas of erythema, mucosal denudation, erosions and ulceration (including kissing ulcers), esophageal blebs, esophageal ulcers with bleeding, or esophageal stricture. In our series, these lesions were most often found in the middle third of the esophagus.

Most of the patients in our study group had developed pill-induced esophageal injury as a result of taking NSAIDs, in particular aspirin. NSAIDs can cause severe esophageal ulceration, particularly in the middle third of esophagus, often associated with bleeding. In a search of the literature we found only 77 cases of NSAID-induced esophageal injury and aspirin was documented as the cause in fewer than 10 % of these. However, aspirin and other NSAIDs accounted for nearly 50 % of all hemorrhagic complications in pill-induced esophageal injury [1]. Considering the frequency with which NSAIDs are prescribed, it appears that NSAIDs are not a common cause of esophageal injury, but when NSAID-induced injury does occur, it is often complicated by hemorrhage [7], as was also observed in the present series. A number of other devastating esophageal injuries have been reported in association with NSAID use, including an ulcer requiring surgical intervention [8], fatal esophageal hemorrhage due to indometacin [9], the development of esophageal strictures in a series of five patients [10], esophageal perforation in one patient [11], and the development of multiple esophageal septa in another [12].

Potassium chloride, alendronate and quinidine were other common medicines which caused pill-induced esophageal injury in our series. Potassium chloride tablets have been reported to cause stricture, presenting with dysphagia, but we did not encounter this in the present series [1]. Alendronate is particularly important because of its relatively common use for the prevention of osteoporosis in postmenopausal women. In one study it was found that 1.3 % of patients who were prescribed alendronate developed esophageal symptoms [12]. In the present series, eight patients had alendronate-related injury, including esophageal stricture in two patients, which was similar to the findings in other studies. [13]. The possibility of Boerhaave's syndrome, which has been reported in association with some pills [14] [15], was excluded in all our patients after reviewing the routine chest radiographs in our series (which has a diagnostic yield of 90 % for esophageal perforation) [16].

Quinidine is notorious for causing extensive exudates, as reported previously [2] [17] [18] [19]. Three patients in our series had quinidine-related esophageal ulceration with lots of exudates in the mid-esophagus.

Various antimicrobial agents have been reported to cause pill-induced esophageal injury, most commonly tetracyclines [1] [20] [21] [22] and penicillins [1]. In this report, 27 patients had esophageal injury due to antibiotics, predominantly tetracyclines. Tetracycline induces injury by accumulating inside the cells and causing inhibition of protein synthesis [2]; it is also acidic when dissolved in water.

The majority of our patients improved by avoiding further exposure to the culprit pill, along with treatment with oral sucralfate and/or proton pump inhibitors. Some of the patients, however, required therapeutic intervention, such as local injection of epinephrine 1 : 10 000 in actively bleeding patients, blood transfusion, and hospitalization. Two of our patients died, showing that pill-induced esophageal injury can be a life-threatening complication. Both patients were elderly and had ischemic heart disease; they had torrential bleeding, a complicated clinical course, and they eventually developed aspiration pneumonia.

The commonest site of esophageal injury was the middle third of the esophagus. It is therefore important that patients with acute esophageal symptoms and a proximal lesion seen on endoscopy should be questioned about medications as a potential cause.

There is a need to create an awareness among healthcare professionals of the caustic nature of certain medicines and of the potential of these medicines to cause serious esophageal injury. These distressing and sometimes life-threatening side effects can be prevented by taking simple precautions, especially in high-risk patients.

References

• 1 Kikendall J W. Pill esophagitis.  J Clin Gastroenterol. 1999;  28 298-305
  PubMedSearch in Google Scholar
• 2 Kikendall J W. Pill-induced esophageal injury. In: Castell DO, Richter JE, (eds.) The esophagus. 3rd edn. Philadelphia; Lippincott Williams & Wilkins 1999: 527-537
  Search in Google Scholar
• 3 Carlborg B, Kumlein A, Olsson H. Drug-induced esophageal strictures [in Swedish, no abstract].  Lakartidiningen. 1978;  75 4609-4611
  PubMedSearch in Google Scholar
• 4 Bonavina L, DeMeester T R, McChesney L. et al . Drug-induced esophageal strictures.  Ann Surg. 1987;  206 173-183
  PubMedSearch in Google Scholar
• 5 Hey H, Jorgensen F, Sorensen K. et al . Esophageal transit of six commonly used tablets and capsules.  BMJ. 1982;  285 1717-1719
  PubMedSearch in Google Scholar
• 6 Semble E L, Wu W C, Castell D O. Nonsteroidal anti-inflammatory drugs and esophageal injury.  Semin Arthritis Rheum. 1989;  19 99
  CrossrefPubMedSearch in Google Scholar
• 7 Kahn L H, Chen M, Eaton R. Over-the-counter naproxen sodium and esophageal injury.  Ann Intern Med. 1997;  126 1006
  PubMedSearch in Google Scholar
• 8 Schreiber J B, Covington J A. Aspirin-induced esophageal hemorrhage.  JAMA. 1988;  259 1647-1648
  CrossrefPubMedSearch in Google Scholar
• 9 Agdal N. Drug-induced esophageal damage: review and report of a fatal case of indomethacin-induced ulceration [in Danish, no abstract].  Ugeskr Laeger. 1979;  141 3019-3021
  PubMedSearch in Google Scholar
• 10 Ryan J M, Kelsey P, Ryan B M, Mueller P R. Alendronate-induced esophagitis: case report of a recently recognized form of severe esophagitis with esophageal stricture - radiographic features.  Radiology. 1998;  206 389-391
  PubMedSearch in Google Scholar
• 11 Corsi P R, de Aguiar J R, de S Kronfly F. et al . Esophageal injury due to pill ingestion [in Portuguese, abstract in English].  Rev Assoc Med Bras. 1995;  41 360-364
  PubMedSearch in Google Scholar
• 12 McCullough R W, Afzal Z, Saifuddin T N. et al . Pill-induced esophagitis complicated by multiple esophageal septa.  Gastrointest Endosc. 2004;  59 150-152
  CrossrefPubMedSearch in Google Scholar
• 13 de Groen P C, Lubbe D F, Hirsch L J. et al . Esophagitis associated with the use of alendronate.  N Engl J Med. 1996;  335 1016-1021
  CrossrefPubMedSearch in Google Scholar
• 14 Cameron R B. Esophagitis dissecans superficialis and alendronate: case report.  Gastrointest Endosc. 1997;  46 562-563
  CrossrefPubMedSearch in Google Scholar
• 15 Adachi W, Watanabe H, Yazawa K. et al . A case of pill-induced esophagitis with mucosal dissection.  Diagn Ther Endosc. 1998;  4 149-153
  PubMedSearch in Google Scholar
• 16 Younes Z, Johnson D A. The spectrum of spontaneous and iatrogenic esophageal injury: perforations, Mallory-Weiss tears and hematomas.  J Clin Gastroenterol. 1999;  29 306
  CrossrefPubMedSearch in Google Scholar
• 17 Teplick J G, Teplicj S K, Ominsky S H, Haskin M E. Esophagitis caused by oral medication.  Radiology. 1980;  134 23-25
  PubMedSearch in Google Scholar
• 18 Wong R KH, Kikendall J W, Dachman A H. Uuinagulate-induced esophagitis mimicking an esophageal mass.  Ann Intern Med. 1986;  105 62-63
  PubMedSearch in Google Scholar
• 19 Ravich W J, Kashima H, Donner M W. Drug-induced esophagitis simulating esophageal carcinoma.  Dysphagia. 1986;  1 13-18
  PubMedSearch in Google Scholar
• 20 Yap I, Guan R, Kang J Y. et al . Pill-induced esophageal injury.  Singapore Med J. 1993;  34 257-258
  PubMedSearch in Google Scholar
• 21 Baumer F, Kellner R, Neumaier U. Doxycycline-induced ulcerous esophagitis [in German, abstract in English].  Fortschr Med. 1997;  115 26-30
  PubMedSearch in Google Scholar
• 22 Levine M S. Drug-induced disorders of the esophagus.  Abdom Imaging. 1999;  24 3-8
  CrossrefPubMedSearch in Google Scholar

S. Abid, FCPS

Aga Khan University Hospital

Stadium Road · PO Box 3500 · Karachi 74800 · Pakistan ·

Fax: +92-21-4934294

Email: shahab.abid@aku.edu

References

• 1 Kikendall J W. Pill esophagitis.  J Clin Gastroenterol. 1999;  28 298-305
  PubMedSearch in Google Scholar
• 2 Kikendall J W. Pill-induced esophageal injury. In: Castell DO, Richter JE, (eds.) The esophagus. 3rd edn. Philadelphia; Lippincott Williams & Wilkins 1999: 527-537
  Search in Google Scholar
• 3 Carlborg B, Kumlein A, Olsson H. Drug-induced esophageal strictures [in Swedish, no abstract].  Lakartidiningen. 1978;  75 4609-4611
  PubMedSearch in Google Scholar
• 4 Bonavina L, DeMeester T R, McChesney L. et al . Drug-induced esophageal strictures.  Ann Surg. 1987;  206 173-183
  PubMedSearch in Google Scholar
• 5 Hey H, Jorgensen F, Sorensen K. et al . Esophageal transit of six commonly used tablets and capsules.  BMJ. 1982;  285 1717-1719
  PubMedSearch in Google Scholar
• 6 Semble E L, Wu W C, Castell D O. Nonsteroidal anti-inflammatory drugs and esophageal injury.  Semin Arthritis Rheum. 1989;  19 99
  CrossrefPubMedSearch in Google Scholar
• 7 Kahn L H, Chen M, Eaton R. Over-the-counter naproxen sodium and esophageal injury.  Ann Intern Med. 1997;  126 1006
  PubMedSearch in Google Scholar
• 8 Schreiber J B, Covington J A. Aspirin-induced esophageal hemorrhage.  JAMA. 1988;  259 1647-1648
  CrossrefPubMedSearch in Google Scholar
• 9 Agdal N. Drug-induced esophageal damage: review and report of a fatal case of indomethacin-induced ulceration [in Danish, no abstract].  Ugeskr Laeger. 1979;  141 3019-3021
  PubMedSearch in Google Scholar
• 10 Ryan J M, Kelsey P, Ryan B M, Mueller P R. Alendronate-induced esophagitis: case report of a recently recognized form of severe esophagitis with esophageal stricture - radiographic features.  Radiology. 1998;  206 389-391
  PubMedSearch in Google Scholar
• 11 Corsi P R, de Aguiar J R, de S Kronfly F. et al . Esophageal injury due to pill ingestion [in Portuguese, abstract in English].  Rev Assoc Med Bras. 1995;  41 360-364
  PubMedSearch in Google Scholar
• 12 McCullough R W, Afzal Z, Saifuddin T N. et al . Pill-induced esophagitis complicated by multiple esophageal septa.  Gastrointest Endosc. 2004;  59 150-152
  CrossrefPubMedSearch in Google Scholar
• 13 de Groen P C, Lubbe D F, Hirsch L J. et al . Esophagitis associated with the use of alendronate.  N Engl J Med. 1996;  335 1016-1021
  CrossrefPubMedSearch in Google Scholar
• 14 Cameron R B. Esophagitis dissecans superficialis and alendronate: case report.  Gastrointest Endosc. 1997;  46 562-563
  CrossrefPubMedSearch in Google Scholar
• 15 Adachi W, Watanabe H, Yazawa K. et al . A case of pill-induced esophagitis with mucosal dissection.  Diagn Ther Endosc. 1998;  4 149-153
  PubMedSearch in Google Scholar
• 16 Younes Z, Johnson D A. The spectrum of spontaneous and iatrogenic esophageal injury: perforations, Mallory-Weiss tears and hematomas.  J Clin Gastroenterol. 1999;  29 306
  CrossrefPubMedSearch in Google Scholar
• 17 Teplick J G, Teplicj S K, Ominsky S H, Haskin M E. Esophagitis caused by oral medication.  Radiology. 1980;  134 23-25
  PubMedSearch in Google Scholar
• 18 Wong R KH, Kikendall J W, Dachman A H. Uuinagulate-induced esophagitis mimicking an esophageal mass.  Ann Intern Med. 1986;  105 62-63
  PubMedSearch in Google Scholar
• 19 Ravich W J, Kashima H, Donner M W. Drug-induced esophagitis simulating esophageal carcinoma.  Dysphagia. 1986;  1 13-18
  PubMedSearch in Google Scholar
• 20 Yap I, Guan R, Kang J Y. et al . Pill-induced esophageal injury.  Singapore Med J. 1993;  34 257-258
  PubMedSearch in Google Scholar
• 21 Baumer F, Kellner R, Neumaier U. Doxycycline-induced ulcerous esophagitis [in German, abstract in English].  Fortschr Med. 1997;  115 26-30
  PubMedSearch in Google Scholar
• 22 Levine M S. Drug-induced disorders of the esophagus.  Abdom Imaging. 1999;  24 3-8
  CrossrefPubMedSearch in Google Scholar

S. Abid, FCPS

Aga Khan University Hospital

Stadium Road · PO Box 3500 · Karachi 74800 · Pakistan ·

Fax: +92-21-4934294

Email: shahab.abid@aku.edu