Recent Research

Hyperglycemia guideline, in-hospital death rates, and more.


Hyperglycemia guideline suggests glucose testing for all inpatients

A recent clinical practice guideline from the Endocrine Society offers recommendations on identifying and managing hyperglycemia in non-critical-care inpatients.

The evidence-based guideline was developed by an expert task force and reviewed by the Endocrine Society, American Diabetes Association, American Heart Association, American Association of Diabetes Educators, European Society of Endocrinology and the Society of Hospital Medicine. The recommendations are intended to offer “practical, achievable, and safe glycemic targets and describe protocols, procedures, and system improvements required to facilitate the achievement of glycemic goals,” according to the guideline.

Eight different subject areas are covered under the guideline, including diagnosis and recognition of hyperglycemia and diabetes, monitoring glycemia, glycemic targets, management of hyperglycemia, recognition and management of hypoglycemia, implementation of glycemic control programs, patient and professional education, and special situations (including surgery and glucocorticoids).

The guidelines' conclusions include:

  • A suggestion that all patients have laboratory blood glucose testing on admission (very low-quality evidence)
  • A recommendation that non-diabetic patients with a blood glucose over 140 mg/dL be monitored with bedside point of care testing for at least 24 to 48 hours (very low-quality evidence)
  • A recommendation for the majority of patients that blood glucose targets be under 140 mg/dL premeal and under 180 mg/dL at random testing (low-quality evidence)
  • A suggestion to discontinue oral hypoglycemic agents and initiate insulin therapy for the majority of patients with type 2 diabetes at the time of hospital admission for an acute illness (very low-quality evidence)
  • A recommendation of insulin therapy as the preferred method for glycemic control in hospitalized patients (low-quality evidence)
  • A suggestion to avoid prolonged use of sliding scale insulin in patients with diabetes (very low-quality evidence)
  • A recommendation to give scheduled subcutaneous insulin therapy consisting of basal or intermediate-acting insulin once or twice a day in combination with rapid- or short-acting insulin administered before meals in patients who are eating (moderate-quality evidence)
  • A recommendation to discontinue oral and non-insulin injectable antidiabetics before surgery (very low-quality evidence)
  • A recommendation for implementation of a standardized nurse-initiated hypoglycemia treatment protocol, initiated when blood glucose is less than 70 mg/dL (low-quality evidence)

The clinical practice guideline was published in the February Journal of Clinical Endocrinology and Metabolism.

In-hospital death rates paint different quality picture than 30-day death rates

Using in-hospital mortality rates to gauge hospital performance favors hospitals with shorter lengths of stay and provides a different picture than 30-day mortality rates, a recent study found.

Researchers used data from the Centers for Medicare and Medicaid Services on patients who were discharged from nonfederal acute care hospitals from 2004 to 2006, and who had complete claims history for the 12 months prior to admission. They created separate cohorts for acute myocardial infarction or AMI (718,508 admissions to 3,135 hospitals), heart failure or HF (1,315,845 admissions to 4,209 hospitals) and pneumonia (1,415,237 admissions to 4,498 hospitals). For patients with pneumonia or HF who had multiple hospitalizations during the study period, only one per patient per year was randomly selected. Patients were excluded if they stayed in the hospital for a day or less, were discharged against medical advice, used Medicare hospice before admission, or had unclear mortality status (i.e., a conflict about the date of death in hospital claims vs. enrollment sources). Outcomes were in-hospital and 30-day all-cause mortality. Results were published Jan. 3 by Annals of Internal Medicine.

Mean patient length of stay (LOS) varied across hospitals for each condition, ranging from 2.3 days to 13.7 days for AMI, 3.5 to 11.9 days for HF, and 3.8 to 14.8 days for pneumonia. The mean in-hospital versus 30-day risk-standardized mortality rates (RSMRs) were 10.8% versus 16.1% for AMI, 5.2% versus 11.2% for HF and 6.4% versus 12.2% for pneumonia. The range across hospitals was wide for all three conditions—around 10 percentage points for any given condition. The in-hospital models resulted in different performance classifications for 257 hospitals (8.2%) for AMI, 456 hospitals (10.8%) for HF, and 662 hospitals (14.7%) for pneumonia. For all conditions, more hospitals shifted to less favorable than more favorable performance categories when the in-hospital measure was used instead of the 30-day measure. Hospital mean LOS was positively correlated with in-hospital RSMRs for all three conditions, with the effect greatest for pneumonia.

Since patients at hospitals with longer LOS are observed for longer lengths of time on average, mean hospital LOS is positively associated with in-hospital mortality rates, the authors noted, while hospital transferred-out rates for AMI are negatively associated with in-hospital mortality. Thus, measurement of in-hospital death rates (rather than 30-day mortality rates) favors facilities with shorter LOS and higher transferred-out rates, they noted. These findings suggest that, although the National Quality Forum endorses either type of mortality measure for public reporting of hospital quality, the use of in-hospital death rates should be reassessed, especially in evaluating the success of health care innovations, the authors wrote. Outcomes measures with standardized follow-up periods that are unaffected by LOS or transfers should be preferred over in-hospital measures, they concluded.

A second study also highlighted the importance of length of stay in performance indicators, noting that 11.3% of patients with ST-segment elevation myocardial infarction (STEMI) were readmitted within 30 days. The comparison in the Jan. 4 Journal of the American Medical Association found that, among 17 countries in North America, Europe, Australia and New Zealand, the United States had the highest 30-day readmission rates after STEMI, but the shortest LOS. When country-level LOS adjustments were made, U.S. location no longer predicted 30-day readmission, however. U.S. location also didn't predict in-hospital death or death within 30 days of discharge, the study found.

Clinicians can use lower hemoglobin threshold for transfusion after hip fracture

A lower hemoglobin threshold for blood transfusion after hip fracture surgery yielded similar outcomes as a more liberal threshold in older patients at high risk of cardiac disease, a recent study found.

Researchers randomly assigned 2,016 patients age 50 years and older to a “liberal” transfusion strategy (to maintain a hemoglobin level of at least 10 g/dL) or a “restrictive” strategy (at physician discretion when hemoglobin level fell below 8 g/dL, or symptoms of anemia appeared). All patients had a history of, or risk factors for, cardiovascular disease and a hemoglobin level less than 10 g/dL within three days after hip fracture surgery. Patients were enrolled from mid-2004 to early 2009 and came from 47 clinical sites in the U.S. and Canada. The main outcome was death or inability to walk 10 feet or across a room without human assistance at 60-day follow-up. Results were published in the Dec. 29, 2011 New England Journal of Medicine.

The mean patient age was 81.6 years, with cardiovascular disease present in 62.9%. In the liberal-strategy group, a median of two units of packed red cells was transfused, compared to none in the restrictive-strategy group. Primary outcome rates didn't differ between groups (35.2% in the liberal-strategy group vs. 34.7% in the restrictive-strategy group). Secondary outcomes also didn't differ significantly, with rates of in-hospital acute coronary syndrome or death at 4.3% in the liberal-strategy group versus 5.2% in the restrictive, and mortality rates of 7.6% and 6.6%, respectively, at 60-day follow-up. Other complication rates were also similar in both groups. There was an interaction between transfusion strategy and sex in the liberal-strategy group that suggested higher mortality or inability to walk without human help at 60 days in men but not women.

Noting that patients in the restrictive-strategy group received 65% fewer units of blood than the other group, and more than half received none, the authors wrote that “widespread implementation of this restrictive approach to transfusion in similar patients would greatly reduce blood use.” It's reasonable to withhold transfusion in post-surgery patients who lack anemia symptoms or who haven't seen a decline in hemoglobin below 8 g/dL, even elderly patients with cardiac disease or risks, they concluded. Editorialists agreed, but warned clinicians to be alert for the risks of undertransfusion, adding that the decision to transfuse should be based on assessment of an individual's signs, symptoms and lab measures, not just the hemoglobin level.

C. diff drugs equally effective for initial cure, review finds

None of the antimicrobial agents used to treat Clostridium difficile (C. diff) in the U.S. is better for initially curing the infection, but recurrence is less frequent with fidaxomicin than vancomycin, according to a systematic review.

Researchers searched MEDLINE, Allied and Complementary Medicine (AMED), ClinicialTrials.gov and Cochrane databases from their inception through August 2011 for randomized, controlled trials in English of adults with C. diff who were treated with medications available in the U.S. They also included less robust studies, such as retrospective cohort studies, that addressed whether treatment effects differed when stratified by disease severity or strain. Two trained extractors abstracted study and patient characteristics and relevant clinical outcomes including initial cure, recurrence, duration of diarrhea and mortality. Results were published Dec. 20, 2011 in Annals of Internal Medicine.

While 1,078 possible relevant articles were found, only 13 (comprising 1,324 patients) met eligibility criteria. Two of these were observational studies that compared the efficacy of different antimicrobials based on C. diff strain, while the remaining 11 trials evaluated the efficacy of different antimicrobials—or doses of one drug—on C. diff infection. Three trials compared metronidazole with vancomycin; eight compared metronidazole or vancomycin with another agent, combined agents, or placebo. Disease strain was analyzed in one trial and two cohort studies.

None of the studies that compared two antimicrobials showed a statistically significant difference in terms of initial cure rates; all these studies were of low-to-moderate-strength evidence. One study of moderate-strength evidence found that infection recurrence was lower with fidaxomicin compared to vancomycin (15% vs. 25%; P=0.005). There was no difference in outcomes of patients treated with vancomycin versus metronidazole, the two most frequently used drugs. Most studies showed no mortality difference between drugs, though a study that compared metronidazole with metronidazole plus rifampin found higher mortality with the latter combination.

Patients with nonsevere disease can be effectively treated with fidaxomicin, metronidazole or vancomycin, the authors concluded, though recurrence—while generally frequent—is less frequent with fidaxomicin compared to vancomycin for non-NAP1 (North American pulse-field gel electrophoresis type 1) strains. While it's commonly believed that vancomycin is superior to metronidazole for severe C. diff disease, there is insufficient evidence of this, they noted. Limitations of the available evidence include that there is large variability among studies, including in the definition of C. diff, initial cure and recurrence, and variability in the duration of treatment and follow up. “In aggregate, these differences make between-study comparisons difficult, and we caution against drawing definitive conclusions,” the authors wrote.

Apixaban not superior to enoxaparin as prophylaxis

In a trial of thromboprophylaxis in medical inpatients, use of apixaban resulted in only slightly lower rates of thrombosis, and more bleeding, than enoxaparin.

The double-blind, double-dummy trial included more than 6,000 inpatients who were acutely medically ill, had an additional risk factor for venous thromboembolism (VTE), and were expected to stay in the hospital for at least three days. They received either oral apixaban, 2.5 mg twice a day for 30 days, or subcutaneous enoxaparin, 40 mg once a day for six to 14 days. The primary outcome was a 30-day composite of death related to VTE, pulmonary embolism, symptomatic deep vein thrombosis (DVT) or asymptomatic proximal-leg DVT, as detected by compression ultrasonography on day 30. The results were published in the Dec. 8, 2011 New England Journal of Medicine.

Of the 2,211 patients analyzed in the apixaban group, 60 (2.71%) experienced one of the primary outcomes, compared to 70 patients (3.06%) in the 2,284-person enoxaparin group (relative risk for apixaban, 0.87; P=0.44). However, there was a significant difference between the groups in major bleeding within 30 days. In the apixaban group, 15 patients had a major bleed (0.47%) compared to six (0.19%) of the enoxaparin patients (relative risk, 2.58; P=0.04). The study authors concluded that an extended course of apixaban was not superior to a shorter course of enoxaparin.

However, they noted several major caveats about the findings. Adverse events increased in the enoxaparin group almost immediately after the drug was stopped. In this time period, when the apixaban group was still receiving therapy, the primary outcome occurred in 31 enoxaparin patients but only 18 apixaban patients. Given the continually increasing rate of events in the enoxaparin group, the study results might have been positive if apixaban had been given for more than 30 days. The authors also noted that the rate of major bleeding with apixaban was less than 0.5%. Given these findings, they said that extended apixaban therapy may still be a promising strategy.

The study also shows that the risk of VTE extends, and even increases, beyond the time of hospital discharge, the authors said. Currently, it is standard practice to stop enoxaparin at discharge, and although this study cannot justify broad policies of extended prophylaxis, it shows the need for more precise risk-stratification strategies to determine which patients would benefit from extended prophylaxis, the authors concluded.

Readmissions frequent, one-quarter related to new problem

Rates and causes of hospital readmissions, and the infrequency of follow-up care, were quantified in a research brief from the National Institute for Health Care Reform.

About a third of adults discharged from the hospital were rehospitalized within one year, according to the brief, which used data from the 2000-2008 Medical Expenditure Panel Survey Household Component. The 5,805 participants, all with at least one overnight hospital stay, underwent five rounds of interviews covering their insurance and health status and health care utilization and expenditures.

Overall, 8.2% of the participants were readmitted within 30 days. More than a quarter of those readmissions (26.1%) were for conditions unrelated to the cause of their initial admission. When the analysis was extended to all readmissions within a year, the percentage of admissions unrelated to the original one was 37.4%. This finding could be explained by patients not receiving adequate treatment for all their comorbidities, the researchers suggested.

The study also found that one-third of the patients did not see an outpatient clinician for follow-up within 30 days after discharge. About 7% of patients visited an emergency department within 30 days of discharge, and this statistic was approximately the same among both patients who did have a follow-up appointment and those who didn't. The study did find that patients who didn't follow up tended to be healthier and younger. There were no significant differences in follow-up rates associated with patients' type of insurance, despite the fact that patients with public insurance were typically sicker, researchers noted.

Private insurance paid for 47% of the 30-day readmissions, while Medicare paid for 40%. This statistic is interesting because more focus has been placed on changing Medicare policy to decrease readmissions, the researchers noted. The authors also discussed how other proposed health care reforms could potentially improve readmission and follow-up rates, including bundled payments, patient-centered medical homes and interoperative health information technology. In addition, they noted that deficits in post-discharge processes are well documented by other researchers, including that less than one-third of outpatient physicians have access to a hospital discharge summary at the time of the first follow up visit. The authors called for additional research on effective care for high-risk patients with multiple comorbidities.

The National Institute for Health Care Reform is a nonprofit organization established by the International Union UAW, Chrysler Group LLC, Ford Motor Company and General Motors that contracts with the Center for Studying Health System Change to conduct health policy research. The brief was released on Dec. 8, 2011.

Infection prevention practices increase, but CAUTI prevention efforts lag

Efforts to prevent hospital-acquired infections generally increased between 2005 and 2009, though prevention practices for catheter-associated urinary tract infections (CAUTIs) were underused, a survey found.

Researchers conducted a longitudinal survey study to compare use of infection prevention practices by U.S hospitals in 2005 and 2009. They compared a random sample of 600 non-federal, general medical and surgical hospitals with 50 or more beds, and 119 VA acute care hospitals, in those two years. They asked infection preventionists at the facilities about use of practices from published guidelines or recommendations meant to prevent central line-associated bloodstream infections (CLABSIs), ventilator-associated pneumonia (VAP) and CAUTI. Respondents were also asked to indicate the perceived effect of rule changes at the Centers for Medicare and Medicaid Services (CMS) on their facility's priority of preventing infections. Results were published online Dec. 6, 2011 by the Journal of General Internal Medicine.

Survey response rate was about 70% in both years. Infection prevention practices improved in 11 of the 12 areas measured, though not always significantly. Ninety percent or more of both non-federal and VA hospitals used three of the recommended practices (maximum barrier precautions and chlorhexidine site antisepsis for CLABSI, and semi-recumbent position for VAP). Other areas of significant improvement in both types of hospitals included use of antimicrobial mouth rinse to prevent VAP and use of portable bladder ultrasound to prevent CAUTI. Areas where prevention practices didn't improve included use of antimicrobial catheters in CLABSI, and use of condom catheters for men as an alternative to indwelling catheters in CAUTI.

More than half of the non-federal hospitals reported a moderate or large increase in infection prevention as a priority due to the CMS non-payment rule. More than 60% of VA hospitals reported no change in priority due to the rule—not surprising, since VA hospitals aren't directly subject to payment change, the authors noted. However, the CMS rule appeared to have limited impact on prevention practices, regardless of survey respondents' perceptions, the authors said. Also during this time period, there was a significant increase in the percent of hospitals reporting they had hospitalists (57% in 2005 and 75% in 2009, P<0.001 at non-federal hospitals; 64% vs. 79%, P=0.016 at VA hospitals).

Despite increases, the use of practices to prevent CAUTI is still relatively low in both types of hospitals, compared to practices to prevent VAP and CLABSI, the authors noted. Prevention practices for the latter infections have been promoted in practice bundles for several years, while there were no such bundles for CAUTI until recently, they said. Also, practices targeting VAP and CLABSI are more likely to affect patients located in the ICU, while those for CAUTI affect patients throughout the hospital, which may make implementation more challenging, they noted. Additional strategies may be needed to encourage use of key CAUTI prevention practices in hospitals, the authors concluded.

Natriuretic peptides may predict outcomes in community-acquired pneumonia

Natriuretic peptides may help predict short- and long-term mortality in patients with community-acquired pneumonia (CAP), a study has found.

Researchers studied 341 unselected patients who presented with CAP to an emergency department in Switzerland between November 2003 and March 2007. N-terminal pro B-type natriuretic peptide (NT-proBNP), mid-regional pro-atrial natriuretic peptide (MR-proANP) and B-type natriuretic peptide (BNP) were measured in a blinded fashion, and their prognostic value for short-term (i.e., 30-day) and long-term mortality was compared with that of two clinical prediction rules, the Pneumonia Severity Index (PSI) and CURB-65 (Confusion, Urea plasma level, Respiratory rate, Blood pressure, and age older than 65 years). Patients were followed for a median of 942 days. The study results were published early online Dec. 1, 2011 by Chest.

Thirty-eight patients (11%) died in the first 30 days, and one hundred fifty-four patients (45%) died during long-term follow-up. Levels of all three natriuretic peptides measured were higher in patients who died in the short and long term than in survivors (P<0.001 for all comparisons). All three natriuretic peptides had prognostic ability similar to that of the PSI and CURB-65 for both short-term and long-term mortality in receiver-operating characteristic analysis. NT-proBNP was found to be an independent predictor of mortality in multivariable Cox regression analysis. The authors combined PSI values and NT-proBNP levels and found that this combination was able to adequately identify patients at low, medium and high risk for death in the short and long term.

The study was limited by its single-center setting, its exclusion of patients receiving chronic renal replacement therapy, and its inability to determine whether effective treatment can be guided by natriuretic peptide testing, since clinicians did not have access to the test results. However, the authors concluded that natriuretic peptides are good predictors of short- and long-term mortality in CAP patients. “A combined assessment using PSI point values and NT-proBNP levels appears to provide additional risk stratification over a single marker approach,” they wrote.