In This Issue...

ACP Sets Best Practice Blood Glucose Target for Intensive Insulin Therapy

A Best Practice Advice paper from the American College of Physicians (ACP) recommends that clinicians target a blood glucose level of 140–200 mg/dL when intensive insulin therapy (IIT) is used in surgical or medical intensive care unit patients (Am J Med Qual 2013; doi:10.1177/1062860613489339). This contrasts with the typical ICU target level of 80–110 mg/dL.

ACP issued the paper as part of its high value care initiative, which promotes the use of diagnostic tests and therapeutic interventions that provide high value, while discouraging the use of low-value tests and interventions that either are not beneficial or are potentially harmful. The panel defined IIT as using intravenous insulin to achieve a targeted blood glucose level with frequent blood glucose testing and adjustment of insulin doses.

The authors noted that well-controlled glucose levels might reduce morbidity and mortality in hospitalized patients. However, studies comparing IIT with normoglycemia in both diabetic and non-diabetic patients in medical or surgical ICUs, and perioperative settings has not consistently shown decreased mortality risk. Results exploring the effects of IIT on length of stay also have been mixed. However, all studies the panel examined found increased risk of hypoglycemia, especially in critically ill patients.

Based on this evidence, the panel concluded that IIT with a goal of achieving normoglycemia or near-normoglycemia does not provide benefits, and might lead to harm. The authors also emphasized that clinicians should avoid targets <140 mg/dL because harms are likely to increase with lower targets.

CMV Levels From Assay Calibrated to WHO Standard Linked to Transplant Outcomes

Solid organ transplant patients with pre-treatment cytomegalovirus (CMV) DNA of <18,200 (4.3log10) IU/mL, as measured by a test calibrated to the World Health Organization (WHO) standard, are 1.5 times more likely to have CMV disease resolution (Clin Infect Dis 2013 56:1546–53). In addition, CMV suppression <137 (2.1log10) IU/mL is predictive of clinical response to antiviral treatment.

CMV disease treatment and viral load monitoring has been shown to improve outcomes in solid organ transplant. However, the field lacks well-defined viral load thresholds because the various quantitative nucleic acid tests for CMV have not been standardized, impeding reproducibility of results. The WHO in 2010 issued an international reference standard for CMV nucleic acid amplification techniques, but until now, no previous clinical studies have been performed using assays calibrated to this standard, according to the authors.

The researchers conducted the study by retesting plasma samples collected as part of an international randomized controlled trial that compared two antiviral medications in patients who had undergone organ transplants. The authors established three predictive rules to determine association between CMV load and outcome, including pre-treatment CMV DNA <18,200 (4.3log10) IU/mL, viral load declines of 1.0, 1.5, and 2.5log10 IU/mL from baseline to days 7, 14, and 21 of treatment, respectively, and viral suppression <137 (2.1log10) IU/mL at days 7, 14, and 21, respectively.

Based on their findings, the authors called for other research teams to conduct clinical studies to determine viral load thresholds for predicting CMV disease risk and for guiding preemptive therapy. A consensus among transplant centers on when to start and stop CMV treatment will emerge as more laboratories calibrate their viral load tests to the WHO CMV standard, the authors suggested.

Fecal Immunochemical Tests Best or Colorectal Cancer Screening

Fecal immunochemical tests (FIT) for hemoglobin outperform guaiac-based fecal occult blood test (gFOBT) across all test characteristics, supporting calls to move to FIT rather than gFOBT in non-invasive colorectal cancer screening programs (Eur J Cancer 2013; doi:10.1016/j.ejca.2013.04.023). The study was one of only a few head-to-head comparisons of FIT and gFOBT in a common screening setting, according to the researchers.

The authors conducted the study because evidence has shown that FITs have higher sensitivity and detection rates of advanced neoplasms, but at the price of a higher positivity rate, which leads to more follow-up colonoscopies. At the same time, gFOBTs have been shown to have high specificity but relatively low sensitivity to detect colorectal cancer. Each test's respective limitations have made it difficult to determine which is better.

The study involved 2,235 participants who provided one stool sample for FIT analysis and who applied stool to two windows of a gFOBT test card prior to receiving a screening colonoscopy, and without dietary or medicinal restrictions in advance of preparation for colonoscopy. The researchers used three enzyme-linked immunosorbent assay-based FITs, and defined each test's cutoff to yield the same positivity rate as the qualitative gFOBT. Any neoplasms were confirmed by screening colonoscopy.

The authors found that all three FITs had substantially better diagnostic performance than gFOBT across all indicators assessed, including sensitivity, specificity, positive-predictive value, negative-predictive value, and positive and negative likelihood ratios. For example, the positive-predictive value for gFOBT to detect any neoplasm was 30.6, versus 64.0, 57.3, and 68.2 for the three respective FITs.

In light of these findings, the authors called for efforts to develop common standards for performing and reporting FITs, including algorithms to define the most appropriate cutoffs for positivity.

Strain-Specific and Biomarker-Associated C. difficile Mortality Described

Clostridium difficile genotype predicts mortality after C. difficile infection (CDI), and excess mortality correlates with genotype-specific changes in biomarkers, strongly implicating inflammatory pathways as a factor in poor outcomes after CDI (Clin Infect Dis 2013;56:1589–1600). The findings not only provide important documentation of C. difficile strain-specific risks for mortality, but also suggest potentially important areas for future investigation, according to the authors.

The researchers conducted the study because some strains, such as polymerase chain reaction (PCR) ribotype 027/NAP1/sequence type (ST) 1, have been associated with not only higher incidence of CDI but also worse outcomes, while others, which now account for most new CDI cases, have not been investigated extensively. In addition, the authors looked at associations between CDI strains and biomarkers, which has not been studied before.

The study involved 2,745 toxin-positive and 27,550 toxin-negative stool samples analyzed by enzyme immunoassay (EIA). Toxin-positive samples also underwent multilocus sequence typing (MLST). The researchers also linked MLST data with hospital admission and discharge data, mortality, and lab test results.

Overall, 14-day mortality was 13% in patients with EIA-positive samples versus 5% with EIA-negative samples. Strain PCR ribotype 078/ST 11 was associated with the highest mortality (25%), followed by PCR ribotype 27/ST 1 (20%). Across a variety of blood or serum markers, the authors also found a significant variance in mean baseline neutrophil counts by genotype, and EIA-positivity versus EIA-negativity. Overall, biomarkers predicted 30–40% of strain-specific mortality differences.