by By Morton Glanz, MD
Nathan RV, Rhew DC, Bratzler DW, et al. In-hospital observation after antibiotic switch in pneumonia: a national evaluation. Am J Med. 2006 Jun;119(6):512.e1-7.
Community-acquired pneumonia (CAP) continues to be a common reason for hospital admission—especially among the elderly. As with many infectious diseases, the duration and route of antibiotic therapy is often based on expert recommendations rather than prospective randomized trials. The Patient Outcome Research Team (PORT) trials address the decision to admit a patient, but not other aspects of care. For hospitalists, the decision of when to discharge any patient with reasonable safety is often fraught with uncertainty. This study addresses the necessity of observing a patient for one day following the switch from IV to oral therapy. Two previous smaller retrospective studies have suggested this was unnecessary.
The current study is also retrospective but involves a large database derived from the U.S. Medicare National Pneumonia Project database. Ultimately 5,248 patients over 65 (mean age=80) were selected for analysis; 2,536 were not observed; and 2,712 were observed for one day.) Patients were excluded if their length of stay was greater than seven days or less than two days, suggesting complicated cases in the former instance and mild illness in the latter (i.e., perhaps not even requiring admission). Immunosupressed patients were also excluded. There was no significant difference in the observed 30-day mortality (5.1% in the “not observed” versus 4.4% in the “observed” cohort, respectively).
The obvious limitation of this study is that it was retrospective/observational and thus potentially subject to the bias inherent in this study design. It is possible that the sicker patients were logically watched longer. Propensity analysis was not a component of this study. The authors do present reasons why certain structural weaknesses would have favored the “observed “group.
Certainly there may be other reasons to observe a patient after the switch to oral therapy. A patient with associated gastrointestinal disturbance or a questionable history of GI or other intolerance to a class of antibiotics is an obvious example. Nevertheless, this study should convey a certain confidence to hospitalists when they assess the suitability for discharge for the type of patient covered in this analysis. Interestingly the recently published guidelines for treatment of community acquired pneumonia are concordant with this study.1
Le Gal G, Righini M, Roy PM, et al. Prediction of pulmonary embolism in the emergency department. Ann Intern Med. 2006 Feb 7:144(3):165-171. Comment in: ACP J Club. 2006 Jul-Aug;145(1):25 & Ann Intern Med. 2006 Feb 7;144(3):210-212.
Pulmonary embolism is a diagnosis frequently considered by the hospitalist—both as an explanation for the admitting clinical picture, as well as a complication arising during the course of a hospitalization for another condition.
My institutions’ ability to identify patients with this potentially lethal condition has greatly improved with the advent of multidetector CT angiography and various diagnostic schemata that include d-dimer testing and estimations of pre-test probability. It is a classic consideration whenever there is a onset of pleurisy, dyspnea, or aggravation thereof. Nevertheless multiple other situations arise in the hospital setting, such as unexplained tachycardia, hemoptysis, or vaguely possible but not clear-cut pleuritic chest pain, in which one feels obligated to at least consider the diagnosis. Further, to have to incorporate d-dimer testing into the diagnostic strategy is problematic as up to 80% of hospitalized patients are likely to be positive. Hospitalists need a reasonable strategy to avoid going down that proverbial pathway in certain low risk situations.
The Geneva scoring system and the Wells system are two methodologies that have been used in lieu of or as an adjunct to “clinical judgment.” The former requires arterial blood gases and the latter has as criteria “other diagnosis more likely than pulmonary embolus” that can be problematic and difficult to standardize.
This article presents a revised Geneva scoring system based solely on elements of the history and physical examination. The elements were derived retrospectively from a prior different study on diagnostic strategies for pulmonary thromboembolism (PTE). A different prospective study on PTE was utilized for the validation arm of this study. By logistical regression analysis the following eight elements were incorporated into the revised Geneva score: Age greater than 65 (1 point), previous deep venous thrombosis or pulmonary embolism (3 points), surgery or fracture within one month (2 points), active malignant condition (2 points), unilateral lower limb pain (3 points), hemoptysis (2 points), heart rate 75 to 94 beats/min (3 points) or heart rate 95 beats /minute or more (5 points), and pain on lower limb palpation and unilateral edema (4 points). The prevalence for pulmonary embolism was as follows: low probability or 8% (0 to 3 points), intermediate probability or 28% (4 to 10 points), and high probability or 74% (equal or greater than 11 points).
Significance for hospitalists: This scoring system is not validated a management system per se. However in the imperfect world of clinical reasoning it can help reinforce a thoughtful decision not to embark on the diagnostic path for pulmonary embolism, with its own inherent risks.
Yang H, Raymer K, Butler R, et al. The effects of perioperative beta-blockade: results of metoprolol after vascular surgery (MaVS) study, a randomized controlled trial. Am Heart J. 2006 Nov;152(5):983-990. Comment in Am Heart J. 2006 Nov;152(5):815-818. McCullough PA. Failure of beta-blockers in the reduction of perioperative events: where did we go wrong? Am Heart J. 2006 Nov;152(5):815-818. Comment in: Am Heart J. 2006 Nov;152(5):983-990.
Hospitalists are frequently consulted regarding perioperative risk assessment and reduction for patients undergoing non-cardiac surgery. Over the last decade and supported by a few studies, the perioperative use of beta-blocker therapy has resolved the uncertainty frequently encountered. The McFalls study in 2004 showed no benefit to routine coronary revascularization for patients undergoing vascular surgery deemed at risk for myocardial ischemia.1 This provided further confidence for those of us supplying these preoperative assessments. However, the Lindenauer study in 2005 (a retrospective cohort analysis) was the first indication that perioperative beta blockade could be harmful.2 Lower-risk patients based on the revised cardiovascular index (RCRI) score actually did worse when treated. Still the ACC guidelines published in 2006 suggested perioperative beta blockers be considered for lower risk patients undergoing vascular surgery.3
This study is a randomized placebo-controlled trial of perioperative beta-blocker therapy in 500 treatment-naïve patients undergoing vascular surgery. Metoprolol was started two hours before surgery and continued for one week. Cardiovascular endpoints included cardiac death, arrhythmia requiring treatment, acute myocardial infarction or acute coronary syndrome, and congestive heart failure. No benefit was found for treatment with metoprolol regardless of the number of Revised Cardiac Risk Index (RCRI) factors present. No excess adverse outcomes were noted for therapy although intraoperative bradycardia and hypotension were significantly increased in the active treatment group.
In the accompanying editorial McCullough discusses possible reasons and implications of these findings. In fact, two other trials have reported similar findings. In contrast to the older trials suggesting a benefit to perioperative beta blockade these newer trials are larger and have a stronger design. He also notes that the patients in the more recent trials are more likely to have prior revascularization and hence are less prone to demand-type events, reflective of the type of insult beta blockade would most likely be helpful in preventing. These events may be more closely allied with plaque destabilization of subcritical lesions, with factors such as perioperative hypercoagulability and perhaps inflammation being more important. In this regard it is notable that recent trials on the perioperative use of statins have demonstrated favorable results, with these agents presumably acting to promote plaque stability, the so-called “pleiotropic” function of statins.
Significance for hospitalists: It is reasonable to be more circumspect in the recommendation of perioperative beta blockade. This practice is not likely the magic bullet, which is a common misconcpetion. An indicative situation is an 80-year-old patient undergoing total hip replacement. He has diabetes, COPD, and hypertension, a pulse of 65, a blood pressure of 110/50. There may also be concerns about bradycardia, hypotension, and bronchospasm. Given this analysis a clinician can be confident in withholding perioperative treatment.
The use of beta-blocker therapy in a patient with a single RCRI factor, which is not coronary artery disease, does not seem justified. On the other hand the use of perioperative statins should be more actively entertained. Emerging recommendations from various specialty organizations and other relevant professional entities should be anticipated and sought.
Bartlett JG. Narrative review: the new epidemic of Clostridium difficile-associated enteric disease. Ann Intern Med. 2006 Nov 21;145(10):758-764.
C. difficile infection is emerging as one of the most important illnesses for hospitalists to be facile with. It not only occurs frequently, but is also often severe or life threatening, and most importantly iatrogenic and preventable. This review by Bartlett, who elucidated the cause of this disease in 1978, reviews important up-to-date information on C. difficile, focusing on the recent emergence of a more virulent form of the disease.
Infectious diarrhea that develops in the hospital is almost always due to C. difficile. The tissue culture cytotoxic assay (first described in 1978) remains the most sensitive and specific diagnostic tool. The toxin immunoassay used most routinely is only 75% sensitive.
An epidemic of unusually severe C. difficile was first described in Quebec in 2001. Important features include a higher tendency for toxic megacolon and a need for colectomy, protein-losing enteropathy, leukemoid reactions, refractoriness to treatment, a high rate of relapse and an astonishing 16.5% attributable mortality. Fluoroquinolones are the leading associated antibiotic causal factor, although extended spectrum cephalosporins remain important as well in this regard. The new strain is characterized by high levels of toxin production due to the deletion of a toxin production regulatory gene. The strain is also fluoroquinolone resistant, explaining the role of that antibiotic in its genesis.
Treatment of C. difficile colitis (especially the emergent strain) remains problematic. In particular the role of metronidazole versus vancomycin as initial therapy is often contentious. Bartlett cites some evidence suggesting vancomycin may be more effective and is especially recommended for severe disease, characteristics of which are often manifested by this new strain.
This review cites important considerations that hospitalist ought to vigilant and proactive in. Given the high risk of fluoroquinolone treatment we must be sure that these drugs are used appropriately. Nonchalantly stacking on levofloxacin therapy for the COPD flair without evidence for pneumonia should be discouraged. When possible antibiotics with a lower risk for C. difficile (sulfonamides, macrolides, tetracyclines) should be used for any infection. When disease is suspected, a negative toxin immunoassay should not discourage empiric treatment especially in a very ill patient. Isolation and barrier precautions are important in preventing the spread of this potentially lethal infection. C. difficile spores are not killed by alcohol-based detergents, and either soap and water or gloves are necessary to care for these patients. When your hospital experiences a clustering of severe C. difficile infection, alert appropriate infection control personnel. Administrative control of antibiotic use may be indicated.
Addolorato G, Leggio L, Abenavoli L, et al. Baclofen in the treatment of alcohol withdrawal syndrome: a comparative study vs diazepam. Am J Med. 2006 Mar;119 (3):276.e13-18.
Alcohol withdrawal syndrome (AWS) is a frequent problem encountered in hospitalized patients; its management is considered one of SHM’s core competencies for hospitalists. Benzodiazepines are the gold standard of therapy for this problem given their established record for safety and efficacy; however, their use can be problematic in certain instances.
This study involved 37 outpatients, as inpatients may often be sicker and at higher risk of severe withdrawal.
There is a definite risk of oversedation—especially in patients with COPD or chronic liver disease. Some patients require inordinately high doses of benzodiazepines, thus setting the stage for a prolonged hospitalization. Occasional paradoxical or disinhibition reactions to benzodiazepines can also be problematic. Addiction and or diversion are also a concern in patients prone to substance abuse. An otherwise stable patient, ready for discharge, may still be on a relatively high dose of lorazepam, but it is generally not prudent to send the patient out with a supply of medication to finish the course given concerns over resumption of drinking while on the sedative. Conversely, the solution can be cold comfort for the attending physician if the patient resumes drinking, thus eliminating the need for additional medication.
Baclofen, a stereoselective gamma-aminobutyric acid agonist, has a long history of safety in the treatment of spasticity. As such it can counter balance the activation of the glutamate excitatory pathway that characterizes AWS. It has been proposed as an alternative treatment for AWS that would not share the above concerns cited for benzodiazepines.
This study is a randomized controlled trial of baclofen versus valium in the treatment for AWS. Thirty-seven subjects with a history of heavy alcohol use were randomized to either baclofen 30 mg per day or valium 0.5 to 0.75 mg/kg. All were outpatients treated for 10 days. Clinical Institute Withdrawal Assessment-Alcohol (CIWA) scores were assessed daily. Both regimens continuously decreased the baseline elevation of CIWA scores daily over the course of the study, without a significant difference in treatment efficacy. No adverse events or side effects were reported in either group.
Other than baseline CIWA and daily alcohol consumption, it is not clear that the two groups were at equal risk for severe withdrawal reactions. Relevant baseline characteristics such as history of seizures or delirium tremens, factors that raise this risk were not noted.
Significance for hospitalists: With a long history of safety and efficacy, benzodiazepines remain the drugs of choice for hospitalists treating patients with AWS. In certain instances it may be desirable to limit or even avoid their use. How effective and safe baclofen would be in filling this role remains to be fully established. In particular the relative risk for sedation and respiratory depression has not been defined. Nevertheless at least in my institutions, as guided by expert consultation, its use has been carefully considered and proven helpful in some of the situations noted above. TH
The Hospitalist newsmagazine reports on issues and trends in hospital medicine. The Hospitalist reaches more than 25,000 hospitalists, physician assistants, nurse practitioners, residents, and medical administrators interested in the practice and business of hospital medicine.