End of Life Care is an emerging field and an important but often overlooked topic to consider in the context of antimicrobial stewardship. When I talk about end of life, I'm targeting the final weeks, days, hours, in a person's life. I'm not focusing on patients with the terminal condition who are still doing well and are expected to remain functional with good quality of life. What is end of life care? There's a general lack of definitional clarity related to several concepts and terms. Terms like palliative care, end of life care, and hospice care, are often used interchangeably, but there are important distinctions that need to be clarified both to patients, their families, as well as clinicians. Palliative care is defined simply as patient and family centered care that optimizes quality of life by anticipating, preventing, and treating suffering. Palliation is the stated goal. Again, the focus is on symptom management especially pain management, not diagnostics, not invasive procedures, not trying to achieve cure. The physical symptoms at the end of life are numerous, and these might include weakness and fatigue, anorexia and cachexia, pain, nausea and vomiting, constipation, confusion, dyspnea, insomnia. It's also interesting that several of these symptoms are side effects that are associated with antimicrobial use. This is an integrated model of palliative care adapted from the World Health Organization. As you can see, early on when patients are healthy, the focus is generally on prevention. Over time, as illness develops, there's a balance both between curative, life-prolonging care, and palliative care. As disease progresses, that balance can shift towards palliative care. In terms of infections, it's not a surprise that patients at the end of life are at increased risk of infection due both to underlying disease as well as disease therapies. Often, these patients can have nutritional reserves that are poor. The threshold to treat possible infection is often quite low and aggressive management with antimicrobials is frequently pursued at times when the overall prognosis is quite poor. So, how often are antimicrobials used at the end of life? My colleagues and I reviewed all patients, most with very advanced disease, seen by the palliative care service at the University of Michigan during a five month period in 2008. We found that about half, 70 of the 131, were receiving antibiotics at the time of initial palliative care consultation. Therapy was empiric in more than half with respiratory infection and urinary tract infection being the most common indications. Pipercillin-tazobactam accounted for 37 percent of antibiotics administered, and this use was empiric in 85 percent of the cases. Vancomycin was given to a third of patients, three-quarters of whom did not have a clear microbiologic indication. Patients managed outside the ICU received antibiotics as frequently as those in the ICU. Overall, our results suggested that broad spectrum antibiotics, almost all empiric, were being used frequently in hospitalized patients near the end of life. We published a follow-up study that examined antimicrobial use at the end of life among hospitalized patients with advanced cancer, by reviewing the records of all patients who died of cancer while hospitalized at the University of Michigan Health System from 2004 to 2007. Among a study population of 145 patients, 26 or 87 percent, received antibiotics for a mean duration of 12.5 days. Of the 126 patients who got antibiotics, 88 or 70 percent had clinical findings suggestive of infection. Although, once again, therapy in more than half was empiric. Seventy two percent received Vancomycin, but only 15 percent of these patients had a clear microbiological indication. Other antibiotics used included Pipercillin-tazobactum, 64 percent, Fluoroquinolones, 65 percent, Cefepime, 24 percent, and Aminoglycosides, 20 percent. C.Difficile complicated six percent of admissions. One one the more remarkable things was that antibiotics were continued in one-third of patients who were placed on comfort care, while overall 45 percent of patients had antibiotic discontinued before death. This occurred on average less than one day prior to death. Based on these observations, we concluded that administration of broad-spectrum antibiotics is frequent, even less than one day prior to death. Our results suggested significant empiric use of broad-spectrum agents among hospitalized patients with advanced cancer even after transition to comfort care. These two studies raised a number of questions. Although we tend to think of antibiotic use as part of usual necessary care in the medical surgical environment, the decision to use or not use antibiotics at the end of life is based on additional factors. Perhaps clinicians feel compelled to offer antimicrobial therapy for possible infection. The reasons for continuing antibiotics especially once the goals of care had clearly transitioned is less clear. So, does antibiotic use at the end of life provide benefits? In order to answer this question, I'd like to look at some of the dementia literature. The next study is a wonderful study done by Erica D'Agata and Susan Mitchell, and was published in 2008. They examined patterns of antimicrobial use among two 214 nursing home residents with advanced dementia. During the time of the study, two-thirds of these patients with very severe cognitive impairment received at least one course of an antibiotic. The mean number per resident was an impressive four courses and a mean of 53 days of antibiotic therapy for a 1,000 resident days was recorded. The most common agents were Quinolones in 38 percent, and third generation Cephalosporins in 15 percent, with the primary indications being pneumonia and urinary tract infection. This is a graph that illustrates the patterns of antimicrobial use during the final eight weeks of patients' lives. As you can see, as patients approach death, the portion of those receiving at least one antimicrobial increased significantly. Besides an increase in the number of courses of antibiotics closer to the time of death, there was also a statistically significant increase in the diagnosis of respiratory infection as an indication for use. In terms of the studies' main conclusions, the authors reported that the number of antimicrobials prescribed in the days of therapy per 1,000 days, all increase significantly as subjects approached death especially during the final two weeks. Some of the same authors published another study that exam survival and comfort after treatment of pneumonia in patients with advanced dementia, who were followed for 18 months or until the time of death. Of the 323 patients included, 133 of 41 percent experienced 225 episodes of suspected pneumonia. This was called suspected because there was no chest x-ray performed in 23 percent. Antibiotics were given to 91 percent with administration by mouth in 55 percent. The authors used standardized scales to assess symptoms and comfort. Antimicrobial administration by any route was associated with lower mortality compared to no antibiotic, and survival at 90 days among those not given an antibiotic was 33 percent. While among those with oral treatment, 64 percent were alive as as were 57 and 61 percent of those given an antibiotic by intramuscular injection and intravenously, respectively. The average adjusted increase in survival time associated with antibiotic administration was 270 days compared to untreated episodes. However, the unadjusted comfort scores were highest for episodes that were not treated with antibiotic agents, and patients had progressively lower scores. In other words, authors observed worse comfort in association with increasingly aggressive care. So, in terms of conclusions, antimicrobial agents were commonly prescribed among patients with advanced dementia as well as pneumonia. Among those who were treated, survival was improved on average by nine months. Oral antibiotics were as effective as intramuscular and IV antibiotics in terms of survival. Among those who received treatment, those patients had lower comfort levels compared to those who received no treatment, and discomfort appeared to increase with more aggressive treatment modalities. The authors offered the following practical clinical guidance. If the goal is to prolong survival even if discomfort is increased, consider limiting therapy to oral or IM modalities, since the benefits appear to be similar. However, if the primary goal is comfort, then antimicrobials should be withheld and palliation provided. The ethics surrounding end of life anti-infective use is an important consideration. Number one, giving antimicrobials at the end of life can delay a patient's transition to hospice. More importantly, patients and families may focus on minor ailments and perhaps even be given false hope, and be distracted from important overall goals. Number two, antibiotics can prolong the dying process. It's an interesting way to look at it. By treating infection, you may, in fact, be extending suffering. Prescribing regimens in congruent with a short life expectancy and goals of care, can increase the reservoir of resistance pathogens and place unreasonable costs on a capitated hospice system. Again, we see that sometimes clinically. Where patients are ready to transition to hospice care. However, they remain on an antibiotic agents which cost a lot of money, have a lot of side effects, and require intensive monitoring. So, how do we translate this into clinical practice? Clinicians often overlook antimicrobial use when addressing other aspects of end of life care. So, obviously, there are some enormous barriers to change in this area. Since it requires a whole new way of thinking not just for clinicians but also patients and their family, current evidence does not support the notion that antibiotics uniformly provide comfort. Now that said, antibiotics are likely more effective in providing comfort for some symptoms than others. A few examples are listed here. Dysuria with urinary tract infection, dysphagia from thrush, diarrhea in the setting of C.Difficile. But for pneumonia, perhaps opioids are a better approach to symptom management. The potential burdens and benefits of treatment with antimicrobials should be carefully assessed in each individual, especially when palliation is a stated goal. Again, going back to the beginning of my presentation, I'm talking about patients who are really in the final few days of their life. Some clinicians may feel uncomfortable even discussing the notion of withholding antibiotics for fear that this may be unethical. Others may believe wrongly that a course of antibiotics is innocuous, but we should keep in mind that antibiotics at the end of life, especially empiric use with IV agents, may distract from the focus of care and delay transition to hospice or even discharged from the acute care setting. In terms of implementing changes in our practice, we can start by reaching out to our colleagues. Start a discussion. Educational sessions are a good place to begin. Consider making antimicrobial use part of advanced care planning and clearly acknowledge futility. More often than not, patients would likely indicate that they would, in fact, want a trial of antibiotics. However, what about not offering antibiotics when the clinical scenario is futile. We have clinicians asking patients and families whether they want this or that, in essence, asking families and patients to make medical decisions instead of having truly informed shared decision making where physicians are giving true guidance and medical recommendations. Again, this should be part of our goals of care discussion. If antimicrobials are used, try to move to oral or even IM formulations if possible. Be mindful of associated side effects and ask yourself and ask the patients, if treatment is contributing to the overall goals of care or simply offering symbolic value. With that, I'm going to end, and I realize I may be leaving you with more questions than answers, but I hope that you will think about end of life care as an area for stewardship interventions.
