Journal watch: Recent studies of note

Oral low-dose steroids for COPD exacerbations, and more.

Oral low-dose steroids yield similar outcomes as high-dose IV steroids for COPD exacerbations

Low doses of steroids administered orally were not associated with worse outcomes than high-dose IV steroids in patients with exacerbations of chronic obstructive pulmonary disease (COPD), according to a study.

Although clinical guidelines recommend low-dose oral steroids for COPD exacerbations, the ideal form and route of administration are unclear. The authors conducted a cohort study of patients admitted to 414 U.S. hospitals with COPD exacerbations in 2006 and 2007, comparing outcomes of those treated with low-dose oral steroids and those treated with high-dose IV steroids during the first two days of their hospitalization. Data were obtained from a fee-supported, voluntary database used to measure quality and health care utilization. Patients who were admitted directly to the intensive care unit were excluded. The authors used a composite measure of treatment failure, defined as mechanical ventilation after the second day of hospitalization, in-hospital death, or readmission for acute COPD exacerbation within 30 days, as the main outcome measure. Secondary outcomes included length of stay and hospital costs. The study results appeared in the June 16, 2010 Journal of the American Medical Association.

Overall, 79,985 patients met the study's enrollment criteria. COPD was the principal diagnosis in 71,628 (90%), and respiratory failure was the principal diagnosis in the remainder (8,357, or 10%). Seventeen percent and 12%, respectively, had been admitted to the hospital for COPD once or at least twice in the preceding year. The authors found that 92% of patients (n=73,765) were treated with IV steroids initially and 8% (n=6,220) initially received oral treatment. In-hospital mortality rates (1.4% [95% CI, 1.3% to 1.5%] vs. 1.0% [95% CI, 0.7% to 1.2%]) and rates of the composite outcome (10.9% [95% CI, 10.7% to 11.1%] vs. 10.3% [95% CI, 9.5% to 11.0%]) were similar regardless of whether IV or oral treatment was received. After multivariable adjustment, risk for treatment failure was not higher in patients treated with low-dose oral steroids than in those who received high-dose IV steroids (odds ratio, 0.93 [95% CI, 0.84 to 1.02]). The authors also performed a propensity-matched analysis and found that orally treated patients had a significantly lower risk for treatment failure, as well as shorter length of stay and lower costs.

The authors noted that their study was observational and that the results should not be generalized to patients with more severe disease who are treated in the intensive care unit. They also noted that physician decisions regarding the choice of therapy may have been influenced by initial disease severity. However, they concluded that low doses of oral steroids were not associated with worse outcomes than high doses of IV steroids in patients with COPD exacerbations, and that “opportunities may exist to improve care by promoting greater use of low-dose steroids given orally.” Physicians may favor high-dose IV steroids over low-dose oral steroids because they may believe high doses are more effective and may lack knowledge of pharmacokinetics, among other factors, the authors wrote.

The authors of an accompanying editorial pointed to the study as a good example of cost-effectiveness research using linked databases and asserted that the data should spur changes in practice. “To ensure that potential benefits supported by observational data are realized, further follow-up evaluations are needed to measure time-trends in quality metrics, health outcomes, and health care costs,” the editorialists concluded.

Hypoglycemia increases mortality risk in pneumonia

Pneumonia patients who are hypoglycemic at hospital admission have a higher mortality risk, creating a red flag that physicians should make sure these patients receive more intensive follow-up and surveillance during hospitalization and after discharge, a study found.

The prospective, population-based Canadian study included almost 3,000 patients who were admitted to six hospitals with community-acquired pneumonia between 2000 and 2002. Patients with a glucose level above 6.1 mmol/L (approximately 110 mg/dL) at admission were excluded and the remaining patients were categorized as normoglycemic (4.0 to 6.1 mmol/L [70 to 110 mg/dL]) or hypoglycemic (less than 4.0 mmol/L). The study appeared in the June 2010 American Journal of Medicine.

In the total study population, 2% of patients were hypoglycemic at admission. Once the hyperglycemic patients were excluded from the study, the percentage increased to 6%. Patients with hypoglycemia were more likely than normoglycemic ones to have severe pneumonia (72% vs. 55%) and to be admitted to the intensive care unit (30% vs. 8%). A higher percentage of them were also diabetic.

The hypoglycemic patients were also more likely to die in the hospital, after 30 days, and after a year, even when factors such as pneumonia severity, ICU admission and having diabetes were controlled for. Their unadjusted risk of mortality was double that of the normoglycemic group both in-hospital (20% vs 9%, P=0.03) and during the next 30 days (20% vs 10%, P<0.01). The increase in risk diminished over time, but remained borderline statistically significant (P=0.05) at one year.

The results show that hypoglycemia is an important and underrecognized prognostic factor in pneumonia, the study authors concluded. The study's findings could also provide new information about the risks of hyperglycemia. Most other research has compared outcomes in hyperglycemic patients to a combined group of normo- and hypoglycemic patients. If hypoglycemic patients have a higher-than-average mortality risk, then the gap between normo- and hyperglycemic patients is likely larger than previously thought.

Little difference between antibiotics in treating acute exacerbation of COPD

Two studies comparing the effectiveness of different antibiotics in treating acute exacerbation of chronic obstructive pulmonary disease (AECOPD) found little difference. Specifically, the studies found:

  • Macrolide and quinolone antibiotics have similar rates of treatment failure in AECOPD, according to a meta-analysis in the May/June 2010 Journal of Hospital Medicine. The analysis of 19,608 patients at 375 acute care hospitals found 69% of patients were treated with quinolone and 31% with macrolide within the first two hospital days. Treatment failure (defined as a composite of inpatient mortality, readmission for AECOPD within 30 days, or starting mechanical ventilation after day two) was the same in both groups in grouped-treatment analysis; there were also no differences in adjusted length of stay or adjusted cost. Antibiotic-associated diarrhea was more common with quinolones (1.2% vs. 0.6%; P=0.0003), as was late mechanical ventilation (1.3% vs. 0.8%; P=0.02), thus “macrolides appear to be the safer of the two,” the authors wrote.
  • Outcomes were similar for patients with severe AECOPD who took ciprofloxacin vs. a combination of trimethoprim and sulfamethoxazole, found a randomized, double-blind trial published in the July 15, 2010 Clinical Infectious Diseases. Of 170 patients with AECOPD requiring mechanical ventilation, 85 received trimethoprim-sulfamethoxazole and 85 received ciprofloxacin for 10 days. The combined outcome of in-hospital death and need for additional antibiotics was similar in the two groups, as was the time to occurrence of the first relapse, duration of mechanical ventilation and length of hospital stay. There were no serious adverse events and tolerance was good overall. The findings highlight the need to reevaluate standard antibiotic therapy regarding its low cost and ability to decrease the use of new, broad-spectrum agents, the authors said.

Duration of ICU delirium associated with long-term cognitive performance

The longer a mechanically ventilated patient has delirium, the worse his or her long-term cognitive performance, a study found.

Researchers studied 126 mechanically ventilated patients in the medical intensive care unit of a large community hospital in Nashville. Trained study personnel assessed patients for delirium every day until discharge, or for a maximum of 28 days, using the Confusion Assessment Method for the ICU (CAM-ICU). Duration of mechanical ventilation was measured from the time of endotracheal intubation to the start of successful unassisted breathing. Cognitive outcomes were assessed at three and 12 months after enrollment by a neuropsychologist who was blinded to the details of each patient's critical illness. Patients were tested using a battery of nine neuropsychological tests designed to measure seven core domains of cognitive functioning, with performance determined by averaging age-adjusted and education-adjusted T-scores from the tests. Results were published in the July 2010 Critical Care Medicine.

Ninety-nine of the 126 patients survived three months or more after their critical illness, and long-term cognitive outcomes were obtained for 78% of patients (n=77). The median duration of delirium was two days. At three-month follow-up, 79% of survivors were cognitively impaired; 62% had severe impairment. At 12 months, 71% of survivors were cognitively impaired; 36% had severe impairment. After adjustment for age, illness severity, severe sepsis, education, preexisting cognitive function, and exposure to sedatives in the ICU, a longer duration of delirium independently predicted worse cognitive performance at three months (P=0.02) and 12 months (P=0.03). Duration of mechanical ventilation wasn't associated with long-term cognitive impairment.

Study limitations include its single-center design and small sample size. The study didn't adjust for depression and post-traumatic stress disorder, which can contribute to cognitive impairment and may confound the relationship between delirium and impairment, the authors noted. Overall, the results highlight the importance of continuing to investigate ways to reduce acute brain dysfunction in the ICU, such as creating protocols for care and certain pharmacologic strategies, the researchers concluded.

Score helps assess risk of adverse drug reactions in elderly patients

A risk score can help identify patients age 65 and older who are at increased risk of an adverse drug reaction, a study found.

In developing their score, Italian researchers tested the GerontoNet ADR (Adverse Drug Reaction) Risk Score on 5,936 patients admitted to community and academic hospitals during survey periods in 1993, 1995 and 1997. The patients were at least age 65 years and received at least one medication during their hospital stay; the mean number of medications taken was 6.3. Of the 5,936 patients, 6.5% (n=383) had an ADR during the hospital stay. Ninety percent of those who had an ADR had only one, 7.8% had two ADRs, and 2.1% had three or more ADRs. Cardiovascular and arrhythmic complications were the most frequent ADRs (25.3% of all ADRs; n=97); then gastrointestinal complications (18%; n=69); then neurologic and neuropsychiatric complications (17.8%; n=68), electrolytic complications (13.1%; n=50), and dermatologic/allergic complications (11.7%; n=45). The strongest predictors for developing ADRs were the number of concurrently used drugs and a history of an ADR, followed by heart failure, liver disease, presence of four or more medical conditions and renal failure. These variables were used to compute the risk score.

The researchers then conducted a validation study on 483 adults admitted to four geriatric and internal medicine wards between September and December 2008. The patients, age 65 years and older, were followed until discharge, and used an average of 11 medications during their stay. An ADR was observed in 11.6% (n =56) of patients. Except for heart failure, all the variables in the risk score were associated with an increased rate of ADRs. An ADR was observed in 28% of patients with a risk score of 8 or more, 11.5% of patients with a score of 6-7, 7.0% of patients with a score of 4-5, 4.2% of patients with a score of 2-3 and 4.5% of patients with a score of 0-1. The area under the curve (AUC), which assesses the ability of the risk score to predict ADRs in this sample, was 0.70 (95% CI, 0.63 to 0.78). Results were published in the July 12, 2010 Archives of Internal Medicine.

Study limitations include that data on preventability of ADRs weren't collected, and that the findings can't be extrapolated to younger persons, the researchers noted. Also, the data used to develop the score were collected between 1993 and 1997; since then, older adult patients have gotten more medically complex. Still, the risk score is a simple, efficient and practical method to identify patients at increased risk of an ADR, the researchers said. ADR incidence in older patients hasn't decreased in the last decade, suggesting a problem with identifying those at risk; the score can help with that, after further validation in different populations and settings, they said.

Use MRI, not CT, to diagnose ischemic stroke, guideline says

A guideline says physicians should use a diffusion MRI scan to diagnose ischemic stroke within 12 hours of a first symptom, instead of a non-contrast CT scan.

The American Academy of Neurology guidelines committee systematically analyzed literature from 1966 to January 2008 on the use of diffusion and perfusion MRI to diagnose stroke. Their other recommendations include:

  • Baseline diffusion MRI volume should be considered useful in predicting baseline clinical stroke severity and final lesion volume in anterior-circulation stroke syndromes (Level B evidence).
  • Baseline diffusion MRI volume may be considered not useful in predicting baseline National Institutes of Health Stroke Scale (NIHSS) score in posterior-circulation stroke syndromes (Level C).
  • Baseline diffusion MRI volume may be considered useful in predicting clinical outcome as measured by the NIHSS and Barthel Index (Level C).
  • Baseline perfusion MRI volume may be considered useful in predicting baseline clinical stroke severity (Level C).
  • There is insufficient evidence to support or refute the value of perfusion MRI in diagnosing acute ischemic stroke (Level U).

The recommendation to use diffusion MRI within 12 hours of a first symptom is based on Level A evidence. The true sensitivity of diffusion MRI for diagnosing ischemic stroke is “probably closer to 80%-90% in a general sample of patients presenting for emergency evaluation of possible stroke,” the guidelines said. False-negative diffusion MRI in ischemic stroke may be due to small strokes, brainstem location and the earliest times from onset, they said. Guidelines were published in the July 13, 2010 Neurology.

Weekend ICU admissions may be associated with higher mortality rates

Patients admitted to the ICU on a weekend may be 8% more likely to die, but overnight admissions face no significantly higher risk, according to a meta-analysis.

Researchers performed a systematic review to examine whether ICU admission at night or on weekends, when less staff are present, is associated with worse outcomes in the critically ill. Studies that evaluated the association between admission time and outcomes and adjusted for disease severity were included, while those that looked at pediatric patients and those not admitted to the ICU were excluded. Overall, the authors examined 10 studies, eight of which looked at nighttime admissions and six of which looked at weekend admissions. The study results appeared in the July 2010 Chest.

Eight studies totaling 135,220 patients evaluated daytime/nighttime admissions. The death rates were 21.4% among 73,676 daytime admissions and 20.8% among 61,544 nighttime admissions. The adjusted odds of death was odds ratio (OR) 1.0 (95% CI, 0.87 to 1.17; P=0.956) between day and night admissions.

A subgroup analysis was done of studies involving intensivist physicians on-site during off-hours. Odds ratios were calculated by study stratification: no intensivist (OR, 1.05; 95% CI, 0.79 to 1.41; P=0.73; heterogeneity P<0.001); intensivist (OR, 0.93; 95% CI, 0.87 to 0.99; P=0.02; heterogeneity P=0.558); and information not available (OR, 1.02; 95% CI, 0.96 to 1.09; P=0.44).

Six studies of 180,600 patients evaluated weekday/weekend admissions. The death rate was 11.1% among 133,150 weekday admissions and 15.6% among 47,450 weekend admissions (OR, 1.08 [95% CI, 1.04 to 1.13]; P<0.001).

A subgroup analysis stratified studies according to the presence of an intensivist physician on-site over the weekend. Again, odds ratios were calculated for death in patients admitted during weekends according to study stratification: no intensivist (OR, 1.07; 95% CI, 0.99 to 1.15; P=0.081; heterogeneity P=0.319); intensivist (OR, 1.03; 95% CI, 0.61 to 1.73; P=0.916); and information not available (OR, 1.09; 95% CI, 1.04 to 1.15; P=0.001).

The authors concluded that weekend ICU admission was linked to 8% higher mortality rates with sampling variability of 4% to 13%, likely related to changes in organizational/staffing structure of ICUs and hospitals over the weekend. Other possible contributing factors include a decreased physician-to-patient ratio, unavailability of board-certified intensivists, physician fatigue, and difficulty in obtaining complex diagnostic tests or therapies. Limitations include that one study dominated the meta-analysis, and that the control group in that study included admissions only Tuesdays through Thursdays.

More patients die from medication errors in July

The rate of fatal medication errors in teaching hospitals spikes almost every July, a finding that can be likely tied to the arrival of new residents, according to a study.

Researchers examined all U.S. death certificates issued between 1979 and 2006, focusing on the almost 250,000 that were attributed to medication errors. The number of deaths observed during the month of July was compared with what would have been expected, based on the deaths that occurred during other months of the year. The study appeared in the July 2010 Journal of General Internal Medicine.

Overall during the time period, and for 21 of the 28 studied years, fatal medication errors spiked by an average of 10% during July. The increase was found only in U.S. counties that had teaching hospitals, and not for any cause of death other than medication errors. The study also found no reduction in the July spike after 2003, when resident work hours were reduced.

The authors looked at a number of alternate explanations before concluding that new medical residents were the most likely cause of the July spike. They noted that the number of inpatient admissions is lower in July, and that any behavioral or societal cause of the medication errors (such as greater alcohol consumption) would affect all counties, rather than just those with teaching hospitals.

Further research is needed to answer remaining questions about the specific mechanisms causing the July spike and to determine why there's no similar increase in surgical errors, the study authors suggested. Still, the results indicate that medical educators should re-evaluate the responsibilities assigned to new residents, increase supervision of them, and provide more education about medication safety, the authors concluded.

Albumin-creatinine ratio in morning urine most predictive

Using a first morning urine sample to measure albumin-creatinine ratio in patients with type 2 diabetes and renal disease offers the best predictor of renal events, according to a study.

Researchers used data from the industry-funded RENAAL (Reduction In Endpoints in Non Insulin Dependent Diabetes Mellitus with the Angiotensin-II Antagonist Losartan) study to compare methods of measuring proteinuria in patients with type 2 diabetes and kidney disease. The four methods tested were urinary protein excretion from a 24-hour urine collection, urinary albumin excretion from a 24-hour urine collection, urinary albumin concentration from a first morning urine sample and albumin-creatinine ratio from a first morning urine sample. The study's primary outcome measure was time to doubling of serum creatinine concentration or time to end-stage renal disease. The results were published Aug. 1, 2010 by the Journal of the American Society of Nephrology.

Seven hundred one patients with type 2 diabetes and nephropathy collected both a first morning urine sample and a 24-hour urine sample. Of these, 202 (28.8%) had a doubled serum creatinine concentration or developed end-stage renal disease during follow-up. Hazard ratios for renal outcome risk per SD increment in each measure were 3.16 (95% CI, 2.60 to 3.86) for urinary albumin excretion, 3.02 (95% CI, 2.53 to 3.62) for urinary protein excretion, 3.23 (95% CI, 2.67 to 3.91) for urinary albumin concentration and 4.36 (95% CI, 3.50 to 5.45) for albumin-creatinine ratio. The albumin-creatinine ratio had a significantly higher area under the receiver-operating curve than the other measures.

The authors cautioned that their results are not generalizable to patients without diabetes and nephropathy, but concluded that measuring albumin-creatinine ratio in first morning urine is the best way to predict progression of renal disease in this population. An accompanying editorial concurred, stating, “Given data from this study and the considerable patient effort required for a 24-hour urine collection, we agree with the authors that the first morning [albumin-creatinine ratio] is in general the logical choice for quantifying proteinuria in clinical practice.”