Acid-suppressive medicines increase pneumonia risk for hospital patients

Ever since a class of drugs called proton pump inhibitors was introduced to the market in the late 1980s, the use of these acid-suppressive medications for heartburn, acid reflux, and other gastrointestinal symptoms has grown tremendously. The widespread use has extended to the inpatient hospital setting, where patients are often routinely given the medications as a way to prevent the development of stress ulcers.

But a large study led by investigators at Beth Israel Deaconess Medical Center (BIDMC) finds that this routine practice may actually be harmful to patients. In their research, published in today’s edition of the Journal of the American Medical Association (JAMA), the researchers found that acid-suppressive medications were associated with a 30 percent increased risk of developing hospital-acquired pneumonia.

“These popular medications are, by and large, considered safe,” explains the study’s lead author, Shoshana Herzig of BIDMC’s Division of General Medicine and Primary Care. “But, in fact, they may be creating an environment that predisposes patients to developing hospital-acquired pneumonia.”

The practice of administering acid-suppressive medications to hospitalized patients originated in intensive care units (ICU), where a combination of stressful conditions and intubation (insertion of a tube into the airway to help a patient breathe) leaves individuals at higher risk of developing stress ulcers. In an attempt to thwart this development, many hospitals began routinely prescribing acid-suppressive agents to all patients in the ICU. And, over time, the practice was also extended to general medical patients, explains Herzig. Today, an estimated 40 percent to 70 percent of hospitalized patients receive some form of acid-suppressive medication, either proton pump inhibitor drugs or medication known as histamine2 receptor antagonists.

“The common thinking has been that these medications are effective at preventing gastrointestinal bleeding, and that they do so without risk,” notes Herzig. She and her co-authors, therefore, decided to determine whether this was indeed the case.

The researchers examined the records of all patients over age 18 admitted to BIDMC and hospitalized for three days or more between January 2004 and December 2007, a total of 63,878 hospital admissions. Using hospital pharmacy records, they determined whether patients were prescribed an acid-suppressing medication during their stay.

Their analysis concluded that acid-suppressive medication was ordered in 32,922 admissions, or 52 percent. (Of the patients who received acid-suppressive medication, 83 percent – 27,236 patients – received proton pump inhibitor drugs, while 23 percent – 7,548 patients – received histamine 2 receptor antagonists, with some exposed to both.) The majority of these medications were ordered within 48 hours of admission.

Using a traditional statistical approach known as multivariable modeling, together with a newer statistical approach using propensity scores, they found that hospital-acquired pneumonia occurred in 2,219 admissions. After further analysis and adjustment for other potential factors that could influence outcomes, the investigators determined that the patients who received the medications had a 30 percent greater risk of developing pneumonia than other patients.

“By reducing acid in the stomach, the medications may be creating a distinct bacterial colonization in the gastrointestinal and respiratory tracts, predisposing patients to develop pneumonia,” explains Herzig. Furthermore, she adds, by reducing the acidity of the stomach contents (as occurs with the use of these medications) a patient’s stimulus to cough may also be reduced, posing another risk factor for pneumonia.

“In conducting this study, our hope is that the use of acid-suppressing medications outside of the ICU would be reserved for patients with clear indications for use, including reflux, gastritis, esophagitis, and identified gastrointestinal bleeding,” notes Herzig.

Study co-authors include BIDMC investigators Edward Marcantonio (senior author), Michael Howell, and Long Ngo. This study was supported, in part, through funding from the Health Resources and Services Administration of the Department of Health and Human Services.