Recent Research

Two questions most effective for initial delirium screening

Asking patients to recite the months of the year backwards was the most effective single-question screen for delirium in older medical inpatients, a recent study found.

The study included 201 inpatients age 75 or older (mean age, 84) treated at a single academic medical center. All of them underwent a clinical reference standard assessment for delirium and the 3-dimensional Confusion Assessment Method (3D-CAM). Researchers compared all the 3D-CAM's items individually and paired to determine which were most sensitive and specific for detecting delirium. Results were published in the October Journal of Hospital Medicine.

Asking patients to recite the months of the year backwards performed better than any other single item, with a sensitivity of 83% (95% CI, 69% to 93%) and a specificity of 69% (95% CI, 61% to 76%). Pairing that question with “What is the day of the week?” resulted in the best 2-item test, with a sensitivity of 93% (95% CI, 81% to 99%) and a specificity of 64% (95% CI, 56% to 70%). The questions worked well even in patients with dementia, and were quick, taking a median of 36.5 seconds in a recent trial, the study authors said.

The results of this study need to be validated prospectively, but if they were, these questions could potentially simplify screening for delirium in hospitalized patients, according to the authors. Positive results on the 1- or 2-question screen would need to be followed up with a more intensive evaluation, such as the full 3D-CAM, since false positives would be common. That could create obstacles to the effectiveness of this system, if hospital staff became reluctant to screen because a positive result would lead to more clinician work, the authors noted.

An accompanying editorial raised a number of other questions about ultrabrief delirium assessments, including the consequences of false negatives and positives, the optimal cutoff on the months of the year backwards test, and whether the assessments will actually improve patient outcomes.

High-sensitivity troponin I with or without advanced cardiac imaging may help assess risk in patients who present with chest pain

High-sensitivity troponin I, with or without computed tomography angiography (CTA), may help stratify risk for cardiac events in chest pain patients, according to 2 recent studies.

In the first cohort study, which was published online Oct. 8 by The Lancet, researchers examined patients with suspected acute coronary syndrome (ACS) at 4 secondary and tertiary care hospitals in Scotland. A high-sensitivity cardiac troponin I assay was used to measure troponin concentration at presentation, and its negative predictive value for index type 1 (myocardial necrosis with suspected acute coronary syndrome) myocardial infarction (MI), or subsequent MI or cardiac death at 30 days (the composite primary outcome) was determined in derivation and validation cohorts.

A total of 6,304 patients were enrolled consecutively in the prospective cohort study; 4,870 patients who had an index MI were included in the derivation cohort. The mean age of the derivation cohort was 64 years, and 43% were women. In the derivation cohort, 782 patients (16%) had an index type 1 MI, 32 (1%) presented again with subsequent MI at 30 days, and 75 (2%) had died of cardiac causes at 30 days.

Most patients without an MI at presentation had troponin concentrations below 5 ng/dL (2,311 of 3,799 patients; 61%), and the negative predictive value was 99.6% (95% CI, 99.3% to 99.8%) for the primary outcome. The researchers also looked at 2 independent validation cohorts and found that 594 of 1,061 patients (56%) without MI at presentation had troponin concentrations below 5 ng/dL; the negative predictive value was 99.4% (95% CI, 98.8% to 99.9%). Patients who had a troponin concentration below 5 ng/dL had lower risk for MI and cardiac death at 1 year than those with a troponin concentration of 5 ng/dL or greater (0.6% vs. 3.3%; adjusted hazard ratio, 0.41; P<0.0001).

The authors noted that they don't know how the troponin threshold used in their study would perform in clinical practice and that they did not know what additional tests and treatments patients received. However, they estimated that based on their results, low plasma troponin concentrations could identify two-thirds of patients who could be discharged from the hospital due to very low risk for cardiac events. “Implementation of this approach could substantially reduce hospital admissions and have major benefits for both patients and health-care providers,” they concluded.

In the second study, a nested cohort observational study of the ROMICAT II (Rule Out Myocardial Infarction/Ischemia Using Computer Assisted Tomography) trial, researchers compared diagnostic accuracy of high-sensitivity troponin I plus advanced coronary artery disease (CAD) assessment with CTA versus conventional troponin and traditional CAD assessment. ROMICAT II included patients who presented to the ED with suspected ACS but had no ischemic changes on electrocardiography and negative conventional troponin results. For the current study, patients with suspected ACS who received coronary CTA and also had high-sensitivity troponin level measured at ED presentation were included. The researchers examined both traditional and advanced features of CAD on coronary CTA. Traditional features were defined as no CAD, nonobstructive CAD, and ≥50% stenosis, while advanced features were defined as ≥50% stenosis and high-risk plaque features. The study results were published online Oct. 13 by JACC: Cardiovascular Imaging.

One hundred sixty patients were included in the current study. The mean age was 53 years, and 40% were women. ACS was diagnosed in 10.6%. All patients had normal levels of conventional troponin at ED presentation. According to traditional coronary CTA assessment, 42.5% of patients had no CAD and were categorized as at low ACS risk, with an ACS rate of 0%; 43.8% had nonobstructive CAD and were classified as at intermediate ACS risk, with an ACS rate of 4.3%; and 13.8% had significant CAD and were categorized as at high ACS risk, with an ACS rate of 72.7%.

In the 9 patients (5.6%) with a high-sensitivity troponin I level below the level of detection, the ACS rate was 0%; for the 139 (86.9%) patients with intermediate high-sensitivity troponin I and the 12 (7.5%) patients whose high-sensitivity troponin I was above the 99th percentile, the ACS rates were 8.6% and 58.3%, respectively. Patients with intermediate high-sensitivity troponin I levels but without ≥50% stenosis or high-risk plaque were able to have ACS ruled out, with a rate of 0%. Those with both ≥50% stenosis and high-risk plaque were found to be at high risk, with an ACS rate of 69.2%, while those with either feature were found to be at intermediate risk, with an ACS rate of 7.7%. Compared with conventional troponin I measurement and traditional coronary CTA, high-sensitivity troponin I and advanced coronary CTA resulted in improved diagnostic accuracy (area under the curve, 0.84 vs. 0.74; P<0.001).

The authors noted that they had high-sensitivity troponin data available for only a subset of the ROMICAT II participants and that their results may not be generalizable to other types of assays, among other limitations. However, they concluded that high-sensitivity troponin I measured at presentation plus early coronary CTA with assessment for high-risk features improved the accuracy of ACS diagnosis versus conventional troponin and traditional coronary CTA assessment. Their results could improve risk stratification of patients who present to the ED with ACS symptoms, they wrote.

Delayed calls for rapid response associated with higher mortality and ICU admission

Activating an emergency team more than 15 minutes after detection and documentation of instability in hospitalized patients was independently associated with an increased risk of ICU admission and death, a study found.

Researchers used data from a 23-hospital cluster randomized trial, the Medical Early Response Intervention Therapy (MERIT) study, which originally studied the impact of introducing medical emergency teams or rapid response teams (RRT) in a randomized, controlled trial. Researchers prospectively collected data from the MERIT study for a new trial that compared hospitals with rapid response teams (intervention group) and hospitals that used standard practice with conventional cardiac arrest team-based responses to emergencies (control group).

A “delayed call” was any call that occurred more than 15 minutes after a criterion for calling the RRT had been documented, including airway at risk, seizure, greater than 2-point decrease in Glasgow Coma Scale (GCS), acute respiratory rate abnormalities, acute pulse rate abnormalities, hypotension, and being subjectively “worried.” Results appeared in the October Critical Care Medicine.

There were 3,135 emergency team calls in all hospitals. Overall, almost one-third of such calls were delayed. Delayed calls were more likely to occur at night, in high dependence or coronary care units, in patients older than 75 years, in those with a decrease in GCS, or in those with hypotension as the reason for the call.

In intervention hospitals, the proportion of delayed calls was similar before and after implementation of rapid response teams. Compared with control hospitals, intervention hospitals had a lower proportion of delayed calls during both the introduction of rapid response teams, (27.3% vs 34.3% weekly rate; incidence rate ratio [IRR], 0.84; P=0.001) and the full implementation period (29.0% vs 34.5% weekly rate; IRR, 0.84; P=0.023).

In all hospitals, delayed calls were associated with an increased risk of unplanned ICU admission (adjusted odds ratio [OR], 1.56; 95% CI, 1.23 to 2.04; P≤0.001) and death (OR, 1.79; 95% CI, 1.43 to 2.27; P<0.001). Delayed calls were also associated with an increased chance of do-not-resuscitate orders being issued after the call.

The authors wrote, “The effect size implies that for every 100 such delays, there are about 13 potentially avoidable deaths. These observations strongly suggest that physicians reappraise the concept of timeliness in the care of such patients and that hospitals consider regularly monitoring the number of delayed RRT activation, ascertain the causes of such delays, and employ appropriate interventions to diminish their occurrence.”

Subclavian catheters associated with fewer bloodstream infections and DVTs than femoral or jugular

Placing a central venous catheter (CVC) in the subclavian vein lowered risk of infection and thrombosis but increased risk of pneumothorax compared to placement in the jugular or femoral vein, a recent study found.

The multicenter French trial randomly assigned the sites of 3,471 nontunneled CVCs in 3,027 adult ICU patients among the subclavian, jugular, and femoral veins (if all 3 sites were appropriate; if not, randomization was done between 2 sites). The primary outcome was a composite of catheter-related bloodstream infection and symptomatic deep vein thrombosis. Results were published in the Sept. 24 New England Journal of Medicine.

The primary outcome occurred in 8 subclavian CVCs, 20 jugular CVCs, and 22 femoral CVCs, for a rate per 1,000 catheter-days of 1.5, 3.6, and 4.6, respectively. The researchers did pairwise comparisons, finding a hazard ratio (HR) for the primary outcome of 3.5 in the femoral group compared to the subclavian group (95% CI, 1.5 to 7.8; P=0.003) and an HR of 2.1 (95% CI, 1.0 to 4.3; P=0.04) for the jugular group compared to the subclavian patients. Between the femoral and jugular groups, risk was not significantly different.

Major mechanical complications occurred most often in the subclavian patients, however, with 18 events in the subclavian group, 12 events in the jugular group, and 6 events in the femoral group. Thirteen of the subclavian complications were pneumothorax, as were 4 in the jugular group. Analysis of these adverse events showed that femoral CVCs were associated with a significantly reduced odds ratio (0.3; 95% CI, 0.1 to 0.8; P=0.03) of mechanical complications compared to subclavian CVCs.

The findings support current CDC recommendations to reduce risk of CVC-related infections by using subclavian sites, the authors noted. However, the overall risk of complications found in the study “suggests that an ideal site for central venous catheter insertion does not exist when all types of complications are considered to be of equal concern,” the authors said. They noted some reasons that the various complications might not be equally concerning, including that the risk of infection and thrombosis increases with the duration of CVC use, while the risk of mechanical complications does not, and that pneumothorax can be diagnosed and treated immediately.

Irrigation of cutaneous abscesses during incision, drainage may not be necessary, study finds

Irrigation during incision and drainage of cutaneous abscesses does not improve 30-day success of treatment, a recent study found.

Researchers randomized 187 patients with cutaneous abscesses to receive irrigation (n=95) or no irrigation (n=92) during their procedures at an academic emergency department. They recorded patient pain during and after the procedures and analyzed each group's need for further invention within 30 days. Results were published online on Sept. 23 by Annals of Emergency Medicine.

In the irrigation group, 14 patients (15%) needed an intervention beyond standard follow-up wound care, compared to 12 patients (13%) in the no-irrigation group, a difference of 2% (95% CI, -8% to 12%). The mean visual analog scale score for procedural pain was also similar between the irrigation and no-irrigation groups (5.6 versus 5.7; difference, 0.1; 95% CI, -0.7 to 0.9).

The groups had similar characteristics, except that the irrigation group was slightly younger (mean age, 35 years versus 40 years) and more often treated with packing (89% vs. 75%; difference, 14%; 95% CI, 4% to 25%) and outpatient antibiotics (91% vs. 73%; difference, 18%; 95% CI, 7% to 29%).

Of 209 patients enrolled in the study, 22 (10 irrigation, 12 no irrigation) were lost to follow-up and not included in the primary data analysis. However, even if all lost no-irrigation patients needed further intervention and the irrigation patients did not, the difference in frequency of intervention would be only 10% between groups (23% versus 13%).

The potential disadvantages of routine irrigation include the need for extra time and supplies, as well as the risk of microbiologic contamination of clinicians and the patient care area, according to the study's authors. They noted that although there may be a perceived benefit to irrigating larger abscesses, the study's 16 patients (4 irrigation and 12 no irrigation) with abscesses greater than or equal to 8 cm had no need for further intervention.

They also noted several limitations of the study, such as how it may not be generalizable to all facilities because they conducted it in a single institution with a high rate of community-acquired methicillin-resistant Staphylococcus aureus. They also noted that patients and physicians were not blinded to the treatment, which could have biased the patients to seek or not seek further care. However, a reviewer who obtained the primary outcome measure was blinded to group assignment.

For moderately ill pneumonia patients, mortality was lower in the ICU than on the ward

Choosing ICU care for pneumonia patients who could instead have been admitted to medical wards appeared to reduce mortality without increasing costs, a recent study found.

The researchers conducted a retrospective cohort study of more than 1 million Medicare beneficiaries admitted for pneumonia between 2010 and 2012. Thirty percent of them were admitted to the ICU, and in adjusted analyses, these patients showed higher 30-day mortality, Medicare spending, and hospital costs than those admitted to general wards. The study was published in the Sept. 22/29 Journal of the American Medical Association.

However, researchers also conducted an instrumental variable analysis, comparing patients who lived within 3.3 miles of a hospital with a high rate of admitting pneumonia patients to the ICU (that is, admitting more than 32%) to those who lived farther away. The likelihood of ICU admission was 36% for the close patients versus 23% for the distant patients (P<0.001). Looking specifically at the 13% of patients whose admission appeared to be based on distance (and thus discretionary), the study found that they had significantly lower adjusted 30-day mortality (14.8% in the ICU vs. 20.5% in the general ward; P=0.02), and no difference in Medicare spending or hospital costs.

The results suggest that, contrary to the study's prespecified hypothesis, admitting patients with uncertain need for ICU care to the ICU improves their survival and doesn't affect costs significantly, the authors concluded. The findings apply only this middle category of patients, not those who definitely do or don't need ICU care, and unfortunately, the study was not structured to identify clinical characteristics defining these patients, they noted. Possible explanations for the findings include ICUs' provision of greater nursing attention, early and aggressive sepsis care, more guideline-based treatment, or care by pulmonary or critical care specialists.

One lesson of the study is that researchers and clinicians should find ways to extend the greatest benefits of ICU pneumonia care (once they are identified) to patients on the general ward, according to an accompanying editorial. The study's findings also “argue against active efforts to reduce ICU admissions through triage guidelines or bed supply reductions,” but that argument shouldn't be taken too far, the editorialists said. “Although it is tempting to use these results as a rationale for more liberal ICU admission strategies, to do so would be untenable in an era of constrained health care resources,” they wrote.