The year in review
Two hospitalists examine the latest evidence that has changed how they practice
Keywords: Recent studies impact physician practice - new research findings on COPD, syncope, and end-of-life counseling
by Katherine Kahn
Published in the December 2009 issue of Today's Hospitalist
Staying up to date on new research—and knowing how to use it—is a constant challenge. At this fall's management of the hospitalized patient conference held at University of California, San Francisco, UCSF hospitalists Bradley Sharpe, MD, and Lindsay Mazotti, MD, detailed the recent studies that have had the biggest impact on how they practice. To highlight findings from more than 15 key studies, Drs. Mazotti and Sharpe presented a series of unfolding cases that contain new evidence on many of the bread-and butter issues—COPD, syncope, end-of-life counseling—that hospitalists encounter every day.
|"The 'shotgun' approach we often use to diagnose and manage syncope is probably not clinically or cost effective."|
–Lindsay Mazotti, MD
University of California, San Francisco
A 60-YEAR-OLD MAN is admitted with dyspnea, increasing cough and pleuritic chest pain. He has a prior history of COPD, coronary artery disease and chronic back pain. Your exam reveals tachypnea and hypoxia.
Q: While it's clear that he is having a COPD exacerbation, should you be worried that a pulmonary embolism (PE) might be an underlying cause? Definitely. A meta-analysis in the March 2009 issue of Chest found that the overall prevalence of PE in patients admitted to the hospital with a COPD exacerbation was as high as 25%.
Unfortunately, researchers found no clear clinical predictors that signaled which patients have a higher PE risk, although some results suggested an association with low PaCO2 or a history of cancer. Although many PEs may not be clinically significant, consider the possibility of PE in all patients with COPD exacerbations, especially those with no apparent inciting factors who are sick enough to be hospitalized.
Q: If diagnostic studies reveal a new unprovoked PE or deep vein thrombosis (DVT), should you be concerned that the patient has cancer?
Yes. A study in the Sept. 2, 2008, Annals of Internal Medicine found that one out of every 10 patients presenting with unprovoked PE or DVT will be diagnosed with cancer within one year—and that up to 6% are diagnosed with cancer at their initial presentation.
"Make sure to do a thorough history, exam and basic lab tests in these patients," said Dr. Sharpe. Ordering a CT scan to detect cancer is likely to be overkill, but take steps to ensure that patients receive thorough primary care follow-up.
Q: You begin standard therapy for COPD exacerbation with bronchodilators, antibiotics and steroids. The patient improves substantially over the next two days and is ready for discharge. How long does someone with a mild to moderate COPD exacerbation need antibiotics?
Five days. An analysis of 21 randomized, controlled trials published in the Jan. 30, 2008, issue of Thorax found no benefit in patients taking antibiotics longer than five days. (Results were the same no matter which antibiotic was prescribed or which setting patients were in, inpatient or outpatient.) Patients with more severe exacerbations likely need longer courses of antibiotics.
Q: Looking over the discharge medication list, you notice that the patient is taking both clopidogrel and a proton pump inhibitor (PPI). Should this drug combination bother you? If so, why?
Possibly. A study in the March 4, 2009, Journal of the American Medical Association found a 25% increase in mortality or recurrent coronary syndrome for patients taking both clopidogrel and a PPI after being diagnosed with acute coronary syndrome (ACS) compared to ACS patients who took only clopidogrel. Researchers found that the risk increased over time.
The study comes on the heels of other research that should lead hospitalists to rethink the liberal inpatient use of PPIs. (See "Research suggests changing how you prescribe PPIs," belo0w.) Dr. Sharpe recommends questioning patients about their PPI use. "Is the patient on a PPI for a clearly defined, evidence-based reason, such as esophagitis or recent ulcer disease?" asked Dr. Sharpe. "If not, you may want to stop it or consider alternative acid suppression."
A 67-YEAR-OLD WOMAN with a history of hypertension and diabetes presents after an episode of syncope.
Q: Which diagnostic tests for syncope will have the highest yield and be the most cost-effective?
Orthostatic vital signs. Research in the July 27, 2009, Archives of Internal Medicine revealed that in older adults with syncope, cardiac enzyme tests, CT scans, echocardiography, carotid ultrasonography and electroencephalography rarely help you figure out the etiology of syncope or affect diagnosis and management. Instead, postural blood pressure readings had the highest diagnostic yield.
In addition, orthostatic vital signs are by far the least expensive option, at a mere $17 per positive test. While electrocardiograms, telemetry and cardiac enzymes are more expensive, they do play some role in diagnosis and medical decision-making.
Select and prioritize tests for syncope based on history and physical exam. "The take-home message is that the 'shotgun' approach we often use to diagnose and manage syncope is probably not clinically or cost effective," Dr. Mazotti said.
Q: After the woman's orthostatics have improved and a UTI diagnosed at admission has been treated, the patient is ready for discharge. You wonder: Why are patients generally discharged late in the day? What can physicians do to improve the discharge process?
Be aware of why discharges are delayed. A prospective study in the April 2009 issue of Journal of Hospital Medicine found that four factors were associated with delayed discharges. Those were:
- use of an ambulance at discharge
- discharge to another health care facility
- filling a prescription at the hospital
- and having a procedure or consult on the final day.
Another study found that many discharge summaries contain significant errors. That research, which appeared in the June 2009 issue of Quality and Safety in Health Care, found that many discharge summaries omit the names of medications patients are taking. Nearly one-third of those omissions could lead to harm.
At the same time, about 17% of all medications listed in discharge summaries had no medical justification. Study results in the September 2009 issue of the Journal of General Internal Medicine showed that discharge summaries mention test results still pending at discharge only 16% of the time. About 10% of those pending results require some action.
One possible solution to improve discharges came in a study in the Feb. 3, 2009, issue of Annals of Internal Medicine. That research found that a team approach to discharge, which included a nurse discharge advocate and follow-up by a pharmacist and primary care provider, resulted in fewer ED visits and 30-day readmissions.
Q: Two months later, the woman returns to the hospital with persistent fevers. You find she has methicillin-resistant Staphylococcus aureus (MRSA) endocarditis, and start her on long-term intravenous vancomycin. Should you calculate her vancomycin dose based on her actual body weight or ideal body weight?
Actual body weight. Guidelines in the Aug. 1, 2009, issue of Clinical Infectious Diseases indicate that vancomycin doses should be based on actual body weight, even if the patient is obese.
The guidelines also state that patients receiving aggressive vancomycin dosing should have their trough concentrations monitored. "Consider weekly monitoring in high-risk patients or those on long-term treatment," said Dr. Mazotti. How frequently patients should be monitored depends on their hemodynamic status, renal function and other factors.
As for what serum trough vancomycin concentration to aim for, the guidelines recommend concentrations greater than 10 mg/L for all patients to be effective and to help avoid resistance. However, in patients with endocarditis, bacteremia, osteomyelitis and other invasive MRSA infections, the goal is 15-20 mg/L.
Q: When should you measure the first trough?
Within 30 minutes before administering the fourth dose. Because that timing is important, don't just order phlebotomy to draw troughs with the morning labs.
Q: On day three, the patient complains of loose stools overnight but reports no abdominal pain. The nurse says she suspects C. difficile. Do you order testing for C. diff toxin?
Not yet. An expert review in the Aug. 4, 2009, issue of Annals of Internal Medicine recommends testing patients for C. diff only if they have three or more loose stools in a day. In addition, a retrospective review in the August 2009 issue of the American Journal of Gastroenterology concluded that repeat tests usually aren't indicated.
That's because 90% of C. diff cases are positive the first time they're tested, and the diagnostic yield on repeat testing drops dramatically. Do repeat testing only in patients with a high pre-test probability of C. diff and, if sending repeat tests, use an alternative diagnostic test if one is available.
AN 89-YEAR-OLD LATINO WOMAN with metastatic colon cancer presents with shortness of breath. She appears frail and confused, and you find that she is in septic shock from pneumonia, requiring pressors and intubation. She is finally stable several hours later, but the ICU nurse tells you the patient's blood sugar is 187 mg/dL.
Q: Should you have the nurse start an insulin drip?
Probably not, as long as the patient's blood glucose remains below 200 mg/dL. That's according to the latest evidence in the ongoing debate over tight glycemic control in critically ill patients.
Published in the March 26, 2009, New England Journal of Medicine, the Normoglycemia in Intensive Care Evaluation–Survival Using Glucose Algorithm Regulation (NICE-SUGAR) trial looked at both medical and surgical patients in the ICU.
Intensive glucose control, which was defined as 81-108 mg/dL, resulted in an increased risk for hypoglycemia and mortality in all patients vs. a more lenient goal of less than 180 mg/dL. Based on all of the evidence available to date, blood glucose levels of between 150 and 180 mg/dL in ICU patients are optimal. (Levels above 200 mg/dL are associated with worse outcomes in all patients, especially those who are critically ill.)
Q: After three days, the patient improves, is transferred to the floor and asks you to remove her urinary catheter. You do so, and decide to order a urinalysis from the urine in her Foley bag. Results show a 104 gram negative rods. Should you start antibiotics for a UTI?
Probably not. The patient more likely has catheter-associated asymptomatic bacteriuria, not an active infection.
Common clinical indicators of infection include pyuria, suprapubic tenderness, fever, urgency or frequency of urinating, dysuria, altered mental status, or hypotension, according to a retrospective analysis in the May 1, 2009, issue of Clinical Infectious Diseases. The study also found that up to one-third of patients inappropriately receive antibiotics for catheter-associated bacteriuria without signs of infection.
Older patients and those with gram negative rods and a higher urine white blood cell count were most likely to have unwarranted antibiotic treatment. "Most hospitalists may be over-treating for catheter- associated UTIs, leading to greater resistance and higher costs," Dr. Sharpe said. "Think before you treat."
Q: Unfortunately, the patient goes on to have a pulseless electrical activity arrest. She is coded for 30 minutes, intubated, given maximum pressors, transferred to the ICU and cooled.
Days pass and she remains comatose with no neurological progress. You decide to meet with her family with the help of an interpreter to discuss limiting life support. When speaking with the family, should you make specific recommendations?
Maybe. Many pulmonary and critical care societies suggest that physicians make recommendations to surrogates about limiting life support. However, according to a prospective study in the Aug. 15, 2009, issue of the American Journal of Respiratory Critical Care Medicine, that may not be the best approach.
Researchers found that the majority of surrogates (56%) wanted to hear physician recommendations. But a substantial percentage—more than 40%—of surrogate decision-makers preferred no physician recommendations for limiting life support.
Katherine Kahn is a freelance health care writer based in western Massachusetts.
Research suggests changing how you prescribe PPIs
A GROWING NUMBER OF STUDIES are delivering the same message: Hospitalists should consider a much more conservative approach to prescribing proton pump inhibitors (PPIs).
Previous studies have documented an association between PPIs and both ventilator-associated pneumonia and community-acquired pneumonia, suggesting these are not as benign a class of drugs as once thought. This year, findings from a study in the May 27, 2009, issue of the Journal of the American Medical Association found an association between PPI use and hospital-acquired pneumonia as well.
Also, a study in the May 2009 issue of the American Journal of Gastroenterology demonstrated that in patients with ascites and cirrhosis, PPI use increased their risk for spontaneous bacterial peritonitis more than four times that of patients not taking PPIs. Moreover, half the patients in that study had no clear indication for being on a PPI in the first place.