Approximately one year later, in preparation for initiating high-dose ibuprofen therapy, the patient's stool was checked for occult blood. He was found to have occasional guaiac positive stools. An upper gastrointestinal endoscopy revealed chronic esophagitis, esophageal ulcerations, and Barrett's esophagus thought to be attributable to gastroesophageal reflux. A 24-hour pH probe revealed significant gastroesophageal reflux despite anti-reflux therapy. Therefore, a fundoplication was performed in order to prevent further esophageal damage. Subsequently, it was believed that his esophageal ulcerations resolved as multiple stools were documented to be guaiac negative.

High-dose ibuprofen therapy was initiated when the patient was 17-years-old, based on published recommended dosage and pharmacokinetic protocols [8]. The patient's dose was determined by a pharmacokinetic analysis [8] that documented a peak ibuprofen plasma concentration of 69 mcg/ml, following a test dose of 1,000 mg (22 mg/kg). A week after initiation of ibuprofen at a dose of 1,000 mg b.i.d. the patient developed severe abdominal pain, hematemesis and bright red blood per rectum. He became hemodynamically compromised and an emergency endoscopy revealed bleeding esophageal ulcerations in the distal 12 cm of his esophagus. After stabilization and observation without further bleeding, a barium swallow demonstrated that his esophagus was deviated towards the right side and the lower segment of the esophagus was relatively horizontal proximal to the gastro-esophageal junction (Figure 1). In addition, a pancreatic enzyme capsule emptied of the enzymes and filled with barium was retained within that esophageal segment for several minutes (Figure 1).

Following this episode the patient's ibuprofen was discontinued, and his oral medications (e.g., antibiotics and vitamins) were switched to liquid form, whenever possible. His pancreatic enzymes were administered after removing the enzyme microspheres from their capsule and mixing them in applesauce.

Reported complications of pneumonectomy include mediastinal shift with herniation of the remaining lung, cardiac herniation, cardiac arrhythmias, bronchopleural fistula, esophageal motility disorders, and development of scoliosis [9-11]. It is thought that pediatric patients have more mediastinal shift following pneumonectomy than adults because of increased elasticity and compliance of the lung and mediastinum that allow for more severe anatomical derangements [9]. A study of 17 post-pneumonectomy pediatric patients revealed that all of the patients had marked herniation of the remaining lung with a mediastinal shift to the opposite side as evident on chest radiographs or computerized tomography scans [11]. One patient with a right pneumonectomy displayed excessive shift of the esophagus as well, but did not have any associated dysphagia or reflux [11]. In another study, esophageal motility was measured in 13 patients before and after pneumonectomy [9]. Patients post-pneumonectomy were shown to have esophageal dysmotility even without reporting dysphagia [9]. The dysmotility was thought to be attributable to the mediastinal shift [9]. Thus, for the patient in the present report, it is likely that the deleterious effects of gastroesophageal reflux may have been increased because of esophageal dysmotility after pneumonectomy.

In our patient, esophageal dysmotility along with the distortion of esophageal anatomy probably combined to slow the esophageal transit time of the ingested ibuprofen, which may have led to development of ulcerations as a result of prolonged concentrated contact within the esophagus. Also, the ibuprofen could have contributed to the development of ulcerations by inhibiting cyclooxygenase systemically, which decreased prostaglandin E production [12]. In turn, less prostaglandin E was available to promote bicarbonate and mucus secretion, which are protective of the gastrointestinal mucosa [12].

Delayed esophageal transit may have occurred with the patient's other oral medications prior initiation of ibuprofen. For example, his pancreatic enzymes likely were retained in the same esophageal segment, which placed him at risk of developing esophageal damage akin to the development of fibrosing colonopathy and strictures described in patients with CF who received high-dose pancreatic enzymes [13]. The enzymes may have remained inactive because the environment was not alkaline enough for their activation [14], which may be the reason that the patient did not demonstrate esophageal strictures.

According to the 2002 CF Foundation registry, only 3.8% of patients in North America with CF were treated with high-dose ibuprofen [15]. In 1999, a survey of 67 CF center physicians revealed that safety issues were a major reason that they did not prescribe this therapy [16]. Gastrointestinal bleeding, a known adverse effect of non-steroidal anti-inflammatory agents, was the safety issue of most concern to these physicians [16]. Other known side-effects of high-dose ibuprofen include renal failure (often transient), and epistaxis [3,17].

The initial randomized, double-blind, placebo control study of high-dose ibuprofen in CF did not demonstrate serious side effects during 4 years of therapy with ibuprofen sufficient to achieve peak plasma concentrations of 50–100 mcg/L [8]. Seven of 85 study patients developed abdominal pain, while only 2 of these 7 were on ibuprofen. One patient in the placebo group developed esophagitis. Of note, abdominal pain, which is very common in patients with CF, actually improved in many patients. In another randomized, double-blind, placebo controlled study, involving 19 children with cystic fibrosis; 13 children received sufficient ibuprofen for 26 months to maximum concentrations 48 +/- 17 mcg/ml, but no adverse effects could be attributed to the ibuprofen [18].