In This Edition:
- Catheter ablation for atrial fibrillation
- Noncontrast CT for appendicitis diagnosis
- Effectiveness of whole-leg ultrasound for DVT diagnosis
- Physician localization and nurse-physician communication
- PPI versus H2 blocker in patients with aspirin-related peptic ulcer disease
- Intensive insulin therapy in steroid-induced hyperglycemia associated with septic shock
- S. aureus decolonization and surgical-site infections
- ABCD2 score and risk of subsequent stroke
Clinical question: In patients with paroxysmal atrial fibrillation (AF) unresponsive to initial antiarrhythmic therapy, what is the efficacy of catheter ablation compared with continued attempts at rhythm control?
Background: Although antiarrhythmic drug therapy (ADT) is generally first-line AF therapy, AF recurrence is high and treatment is associated with adverse effects. Catheter ablation is an alternative treatment. Recent studies comparing antiarrhythmic drugs and catheter ablation have involved small populations and have shown mixed results.
Study design: Prospective, multicenter, unblinded, randomized trial.
Setting: Nineteen hospitals—15 in the U.S.—with considerable experience in AF ablation.
Synopsis: This trial compared catheter ablation (n=106) vs. ADT (n=61) for symptomatic, paroxysmal AF refractory to at least one antiarrhythmic drug. Patients in the ablation arm were allowed up to three procedures within an 80-day period; patients in the ADT arm were treated with a previously unused class I or class III antiarrhythmic. After nine months, 34% of ablated patients had failed treatment compared with 84% of patients receiving ADT (HR 0.30 [95% CI, 0.19-0.47]).
While the results are encouraging, some limitations should be noted. Attempts to generalize the results of this study might be limited, as the hospitals had considerable experience in AF ablation, the patient population was relatively young (mean age=56 years), and patients with significant left ventrical dysfunction and persistent AF were excluded.
Furthermore, the long-term effectiveness of ablation was not evaluated, and the study did not assess such outcomes as mortality, stroke, or AF progression. The effectiveness of specific ablation techniques could not be determined because a variety of approaches was employed.
Bottom line: Patients with symptomatic paroxysmal AF without advanced heart failure might benefit from catheter ablation at experienced institutions.
Citation: Wilber DJ, Pappone C, Neuzil P, et al. Comparison of antiarrhythmic drug therapy and radiofrequency catheter ablation in patients with paroxysmal atrial fibrillation: a randomized controlled trial. JAMA. 2010;303(4):333-340.
Clinical question: What is the diagnostic accuracy of noncontrast helical CT in the evaluation of suspected acute appendicitis in the ED?
Background: In the workup of acute appendicitis, various CT protocols are used, including combinations of oral, intravenous, and rectal contrast, as well as noncontrast protocols. Noncontrast CT is less time-consuming and avoids risk of allergic reaction or contrast-induced nephropathy. The diagnostic accuracy of noncontrast CT, however, is controversial.
Study design: Systematic review/meta-analysis.
Setting: Seven studies evaluating acute appendicitis with noncontrast CT in ED settings.
Synopsis: This is the first systematic review of noncontrast CT (helical/multislice) in adults with suspected appendicitis. The authors pooled seven studies (1,060 patients) comparing noncontrast CT with a reference standard of final diagnosis at surgery or followup at a minimum of two weeks. The review yielded the following pooled estimates: sensitivity 93%, specificity 96%, positive likelihood ratio 24, and negative likelihood ratio 0.08. Overall, the diagnostic accuracy of noncontrast CT was high.
A few limitations should be noted. In the original studies, results were reported inconsistently. In one study, 24% of scans were inconclusive, with an associated likelihood ratio of approximately 1. Future studies must address this important clinical question of how to treat patients with inconclusive scans. Another question is to what degree the accuracy reported in this study reflects the expertise of the institutions, the majority of which were university-affiliated. According to one study author, “a certain level of experience is required for skillful interpretation” of noncontrast CT.
Bottom line: Noncontrast CT has a high sensitivity and specificity for acute appendicitis, and should be considered an alternative to contrast CT, particularly in patients with contraindications to contrast or those at risk of contrast-induced nephropathy.
Citation: Hlibczuk V, Dattaro JA, Jin Z, Falzon L, Brown MD. Diagnostic accuracy of noncontrast computed tomography for appendicitis in adults: a systematic review. Ann Emerg Med. 2010;55(1):51-59.e1.
Clinical question: How useful is compression ultrasound (CUS) for diagnosing distal DVT?
Background: CUS can reliably evaluate proximal DVT, but its accuracy for distal DVT is controversial. Because of the risk of extension of distal DVT (up to 25%), guidelines recommend that some patients undergo serial proximal CUS after an initial negative result. As an alternative, recent studies have evaluated one-time, whole-leg CUS.
Study design: Systematic review and meta-analysis.
Setting: Review of randomized controlled trials and prospective cohort studies.
Synopsis: The study pooled data from seven studies and more than 4,700 patients with suspected DVT for whom anticoagulation was withheld after a single, negative, whole-leg CUS. At the three-month followup, the combined symptomatic VTE event rate was 0.57%, and the authors concluded that withholding anticoagulation was associated with a low VTE risk.
Although encouraging, this study had several limitations. First, whole-leg CUS is not widely performed or standardized, and the CUS technique varied slightly across the studies. Second, any attempt to generalize the results of this study might be limited, as the proportion of pregnant patients and those with malignancy was low. Furthermore, only one of the seven studies included inpatients that might be at higher VTE risk. Third, pre-test probability was assessed for only a subset of patients, limiting assessment of VTE by risk level. For example, the overall finding of the study—an event rate of 0.57%—appears low but is difficult to apply clinically when subset analyses for high-risk patients yielded a VTE rate of 2.5% with wide confidence intervals ranging from 0% to 7%.
Bottom line: Whole-leg CUS might be a practical alternative to serial proximal CUS, but more data incorporating pre-test probabilities and involving more inpatients are needed. Hospitalists should be cautious in applying pooled summary estimates.
Citation: Johnson SA, Stevens SM, Woller SC, et al. Risk of deep vein thrombosis following a single negative whole-leg compression ultrasound: a systematic review and meta-analysis. JAMA. 2010;303(5):438-445.
Clinical question: Does localization of inpatient physicians on hospital units improve communication with nursing staff?
Background: While nurses are typically localized on a hospital unit, inpatient physicians often care for patients on multiple units. This lack of regionalization makes it difficult for physicians and nurses to discuss care plans directly. No prior research has evaluated the effect of physician localization on nurse-physician communication.
Study design: Cross-sectional, pre- and postintervention study.
Setting: Tertiary-care teaching hospital, general medical service.
Synopsis: The study was a cross-sectional survey of nurses and physicians prelocalization (n=342 patients) and postlocalization (n=294 patients) of physicians on hospital units. Localization was associated with increased frequency of communication; however, it did not improve the consistency of nurse-physician agreement on the care plan. Nurse-physician agreement was improved on two aspects of the care plan—planned tests and anticipated length of stay—but not on primary diagnosis, planned procedures, medication changes, or consultations.
Limitations of the study were that it was conducted at a single teaching hospital, communication patterns might have changed during the year between pre- and postlocalization, and physicians were not completely localized to specific units (73% localization).
Despite the limitations, this study is the first to evaluate staff localization and communication on a general medical service. The findings suggest that localization is a first step toward interdisciplinary communication. It also shows that quality and content of communication require further assessment.
Future studies must assess the impact of communication on the quality of patient care.
Bottom line: Physician localization improved the frequency of nurse-physician dialogue but did not consistently facilitate a shared understanding of the care plan. Although not assessed in this study, the implication is that the quality of communication between providers needs improvement.
Citation: O’Leary KJ, Wayne DB, Landler MP, et al. J Gen Int Med. 2009;24(11):1223-1227.
Clinical question: Is a twice-daily H2-receptor antagonist (H2RA) or a once-daily proton pump inhibitor (PPI) better in patients who must continue low-dose-aspirin therapy despite aspirin-related peptic ulcer disease?
Background: Some patients with aspirin-related peptic ulcer disease require continued aspirin therapy. It often is assumed that PPIs are superior to H2RAs in secondary prevention of low-dose aspirin-related injury, although no randomized trials have specifically addressed this question.
Study design: A prospective, double-blind, randomized controlled trial.
Setting: A university hospital in Hong Kong.
Synopsis: In this trial, 160 patients with aspirin-related peptic ulcers/erosions were randomized to 48 weeks of oral famotidine (40 mg twice daily) or pantoprazole (20 mg daily) after mucosal healing and eradication of H pylori. During this time, all patients continued to receive aspirin (80 mg daily).
The rates of recurrent dyspeptic or bleeding ulcers/erosions within 48 weeks were 20% with famotidine versus 0% with pantoprazole. The rates of gastrointestinal (GI) bleeding alone were 7.7% versus 0%, respectively.
Of note, none of the five patients with GI bleeding had significant dyspepsia, which is consistent with prior reports that NSAID-induced injury might be silent.
Overall, high-dose famotidine was inferior to pantoprazole in the prevention of recurrent aspirin-related injury. These findings support the ACCF/ACG/AHA 2008 recommendation of PPIs over double-dose H2RAs in this clinical setting. It is not known, however, whether the PPI strategy extends to average GI-risk patients taking aspirin.
Furthermore, although not demonstrated in this study, it is possible that the lower rates of dyspepsia with pantoprazole might facilitate adherence to prolonged aspirin therapy.
Bottom line: High-dose famotidine was inferior to pantoprazole in the prevention of recurrent low-dose-aspirin-related injury.
Citation: Ng FH, Wong SY, Lam KF, et al. Famotidine is inferior to pantoprazole in preventing recurrence of aspirin-related peptic ulcers or erosions. Gastroenterology. 2010;138
Intensive Insulin Therapy Is Not Beneficial in Corticosteroid-Induced Hyperglycemia Associated with Septic Shock
Clinical question: In septic shock treated with hydrocortisone, does intensive insulin therapy reduce in-hospital mortality compared with conventional glucose management?
Background: Corticosteroids might benefit patients with septic shock, but they also can exacerbate illness-induced hyperglycemia. It is hypothesized that normalization of blood glucose with intensive insulin might improve outcomes in these patients.
Study design: A multicenter, 2x2 factorial, randomized controlled trial.
Setting: Eleven ICUs in France.
Synopsis: In this 2x2 factorial comparison, 509 patients with septic shock treated with hydrocortisone were randomized to IV insulin, conventional insulin, fludrocortisone plus IV insulin, and fludrocortisone plus conventional insulin. The primary objective was to assess intensive IV versus conventional insulin, and a secondary objective was to assess the benefit of adding fludrocortisone to hydrocortisone therapy.
Overall, analysis showed no difference in in-hospital mortality in either of the two comparisons.
These null findings should be interpreted considering the following study limitations. In the intensive-insulin groups, there was a failure to reach target glucose levels of 80 mg/dL to 110 mg/dL (mean achieved levels 120 mg/dL to 130 mg/dL and higher). These levels overlapped to some degree with the control group, which targeted glucose levels <150 mg/dL and achieved mean levels of 140 mg/dL to 150 mg/dL.
The lack of substantial difference in glucose levels might have contributed to the null findings. Furthermore, the absolute risk reduction in the original sample-size calculations was overestimated. The result was an underpowered study, which also might have contributed to the null findings.
Bottom line: In septic shock treated with hydrocortisone, the optimal blood-glucose level and insulin strategy are unknown.
Citation: Annane D, Cariou A, Maxime V, et al. Corticosteroid treatment and intensive insulin therapy for septic shock in adults: a randomized controlled trial. JAMA. 2010;303(4):341-348.
Clinical question: Can rapid screening and decolonization of nasal carriers of Staphylococcus aureus on admission reduce surgical-site S. aureus infections?
Background: More than 80% of healthcare-associated (HCA) S. aureus infections are endogenous in origin. High-level nasal carriers have a three- to sixfold increased risk of infection with this organism compared with noncarriers and low-level carriers. Decolonization of nasal and extranasal S. aureus on admission might reduce this risk of infection.
Study design: Randomized, double-blinded, placebo-controlled, multicenter trial.
Setting: Three university hospitals and two general hospitals in the Netherlands.
Synopsis: In this study, 918 mostly surgical patients with nasal S. aureus identified preoperatively by PCR tests were randomized to decolonization versus placebo. The five-day decolonization protocol involved mupirocin nasal ointment and chlorhexidine soap baths. Decolonization reduced length of stay by nearly two days. Through six weeks postdischarge, the cumulative incidence of S. aureus infection was 3.4% after decolonization versus 7.7% with placebo (RR 0.42; 95% CI, 0.23-0.75). Among the sites of infection, deep surgical sites had the greatest risk reduction (RR 0.21; 95% CI, 0.07-0.62).
The results of this study are encouraging, but a few limitations should be noted. The decolonization protocol lasted five days, which might make implementation less practical. Also, the relative contributions of mupirocin and chlorhexidine are unclear.
S. aureus is important, but it represents a minority of surgical-site infections; the effect of the protocol on other organisms is unknown. Lastly, MRSA is not prevalent in the Netherlands and no carriers were identified in the study. Although the protocol was designed to eradicate MRSA, such carriers might have different carriage patterns requiring throat swabs in addition to nasal swabs.
Bottom line: Preoperative detection of S. aureus nasal carriage and nasal and extranasal decolonization significantly reduced endogenous S. aureus infection and length of stay. Decolonization might be most beneficial for carriers at increased risk of deep infection, such as those undergoing cardiac surgery.
Citation: Bode LG, Kluytmans JA, Wertheim HF, et al. Preventing surgical-site infections in nasal carriers of Staphylococcus aureus. N Engl J Med. 2010;362(1):9-17.
ABCD2 Is a Poor Predictor of Early Ischemic Stroke after Transient Ischemic Attack
Clinical question: How well does the ABCD2 score predict stroke risk within seven days of transient ischemic attack (TIA)?
Background: After TIA diagnosis, the seven-day risk of ischemic stroke is approximately 5%. Identifying these high-risk patients might facilitate ED decision-making. The ABCD2 score, a prediction tool for stroke after TIA, might be useful but has not been prospectively validated in a large, broad-patient population.
Study design: Prospective validation study, convenience sample.
Setting: Sixteen hospitals in North Carolina.
Synopsis: The ABCD2 score (range 0-7 points) predicts stroke risk after TIA. The investigators evaluated the accuracy of ABCD2 in predicting seven-day ischemic stroke risk in a convenience sample of 1,667 TIA patients. Strokes were categorized as disabling or nondisabling.
Overall, the score was poorly predictive of all ischemic stroke (c stat 0.59) and moderately predictive of the subset of disabling ischemic stroke (c stat 0.71). The ABCD2 had the most discriminatory power when used to identify patients at low risk of disabling stroke (0-3 points); for these patients, the negative likelihood ratio (LR) was 0.16 (0.04-0.64).
The study is the largest published external-validation study of the ABCD2 score, but it had significant limitations that should be considered. There was potential sampling bias because of nonconsecutive sampling, and unaccounted patients with TIA were discharged from the ED. Furthermore, ABCD2 scores were incalculable for 35% of patients, although the authors report that imputed data did not change the findings significantly.
Bottom line: The ABCD2 score does not sufficiently predict the seven-day risk of ischemic stroke after TIA. Further validation studies are needed.
Citation: Asimos AW, Johnson AM, Rosamond WD, et al. A multicenter evaluation of the ABCD2 score’s accuracy for predicting early ischemic stroke in admitted patients with transient ischemic attack. Ann Emerg Med. 2010;55(2):201-210.e5. TH
Dr. Kim is a hospitalist at Brigham and Women’s Hospital in Boston, and an instructor at Harvard Medical School.