Recurrent C. diff, decision tree for sepsis, and more

Summaries from ACP Hospitalist Weekly.

Recurrent C. difficile infections more than doubled between 2001 and 2012

The incidence of multiply recurrent Clostridium difficile infection increased dramatically between 2001 and 2012, according to a recent study.

The retrospective cohort study used data on 38 million commercially insured U.S. patients, of whom 45,341 developed C. difficile. Of those patients, 1,669 had multiply recurrent C. difficile. Between 2001 and 2012, the annual incidence of C. difficile increased by 42.7% (from 0.4408 to 0.6289 per 1,000 person-years). In the same period, incidence of multiply recurrent infections increased 188.8% (from 0.0107 to 0.0309 per 1,000 person-years). Results were published online July 4 by Annals of Internal Medicine and appeared in the Aug. 1 print issue.

Researchers also looked at risk factors for recurrent infection. Compared to patients with non-recurring C. difficile, those with multiply recurrent infection were older (median age, 56.0 years vs. 49.0 years), more likely to be female (63.8% vs. 58.7%), and to have used antibiotics (72.3% vs. 58.8%), proton-pump inhibitors (24.6% vs. 18.2%), or corticosteroids (18.3% vs. 13.7%) within 90 days of C. difficile diagnosis. Chronic kidney disease and diagnosis in a nursing home were also associated with risk for multiply recurrent C. difficile infections (10.4% vs. 5.6% and 2.1% vs. 0.6%, respectively).

The goals of the study included identifying risk factors for multiply recurrent C. difficile and determining how many patients would be candidates for fecal microbiota transplantation in coming years, the study authors noted. Based on the results, demand for fecal transplants and new antimicrobial therapies can be expected to increase considerably, they said. The increase over time in the incidence of multiply recurrent C. difficile infections, even after adjustment for risk factors, may be due to a change in the biology of C. difficile, such as the emergence of the North American pulsed-field gel electrophoresis type 1 strain, the authors speculated.

The identified risk factors for recurrence could be used to adjust treatment algorithms for high-risk patients, including earlier use of fecal transplant. Limitations of the study included that the definitive causes of the observed increase could not be determined and could include factors such as increased testing for C. difficile, although the study largely predated widespread adoption of nucleic acid amplification testing.

An accompanying editorial said this is the largest study to date of multiply recurrent C. difficile and its epidemiology. The editorialists noted additional limitations, including that the study did not include fee-for-service Medicare patients and did not capture some factors associated with recurrent infection that could partially account for the observed increase, including prior fluoroquinolone use.

Decision tree can predict resistance to antibiotics commonly used for sepsis

Researchers recently developed a tool for predicting whether sepsis caused by gram-negative bacteria would be resistant to piperacillin-tazobactam (PT), cefepime, or meropenem.

To develop the tool, researchers studied 1,618 consecutive adult patients admitted to a single hospital between 2008 and 2015 with sepsis or septic shock due to bloodstream infections caused by gram-negative bacteria. The study used multivariable logistic regression to identify risk factors for resistance to the three antibiotics and developed clinical trees using a recursive partitioning algorithm. Results were published online by Clinical Infectious Diseases on July 10.

Photo by Thinkstock
Photo by Thinkstock.

The prevalence of antibiotic resistance was 28.6% for PT, 21.8% for cefepime, and 8.5% for meropenem. Prior antibiotic use, nursing home residence, and transfer from an outside hospital were predictors of resistance to all of the drugs. Meropenem resistance was predicted by infection with Pseudomonas or Acinetobacter spp. The decision trees identified clusters of patients with low probability of meropenem resistance and clusters with high probability of resistance to all of the drugs.

For example, central nervous system and central venous catheter infections were associated with resistance to PT, while an unknown source of infection was protective against meropenem resistance. Patients who were admitted from nursing home and had a central venous catheter had a high likelihood of resistance to PT, while cefepime resistance was most common in patients requiring mechanical ventilation and central venous catheter.

The authors concluded that clinical decision trees based on simple variables could identify patients at low, intermediate, and high risk of resistance to these three antibiotics. They noted that previous research has focused on predicting the causative species of bacteria, but that a tool to predict resistance might be more relevant to practicing clinicians. They noted that risk factors might not vary by institution but that their relative weights and interactions “probably do,” and models could be developed and validated locally to predict resistance. However, limitations include that risk factors might vary temporally and that some hospitals might not have the laboratory services to rapidly determine whether an infection is gram-negative.

Decision trees like these should be used to tailor antibiotic treatment, the authors said. Patients at low risk of resistance to PT or cefepime could be treated with those medications, while those at intermediate risk should receive double coverage, possibly with an aminoglycoside. Meropenem should be considered, possibly along with a second agent, for patients at high risk of resistance to the other drugs. Patients at high risk of meropenem resistance should probably receive either two different classes of drugs or a new broader-spectrum antibiotic, the authors advised.

Hospital revisits, admissions appear common after observation stays in Medicare beneficiaries

Observation stays in Medicare beneficiaries are often followed by revisits and inpatient admissions, a recent study indicates.

Researchers conducted a retrospective cohort study at 4,750 U.S. hospitals using data from a nationally representative sample of Medicare fee-for-service beneficiaries 65 years of age or older to determine rates and trends over time of outcomes after an observation stay. The study's main outcome measures were rates of ED treatment-and-discharge stays, repeat observation stays, inpatient stays, any revisit to the hospital, and death within 30 days of discharge after index observation stays. The researchers compared these rates with corresponding outcomes within 30 days of discharge from index ED treatment-and-discharge stays and index inpatient stays. Study results were published online June 20 by BMJ.

Overall, discharges after 363,037 index observation stays, 2,540,000 index ED treatment-and-discharge stays, and 2,667,525 index inpatient stays from 2006 to 2011 were included in the study. After the index observation stays, 30-day rates of ED treatment-and-discharge stays, repeated observation stays, inpatient stays, any hospital revisit, and death were 8.4%, 2.9%, 11.2%, 20.1%, and 1.8%, respectively. From 2006 to 2011, an increase was seen in revisit rates for ED treatment-and-discharge stays, repeated observation stays, and any hospital revisit (P<0.001 for trend). Thirty-day rates of inpatient stays and 30-day mortality rates did not change in this time period, however (P=0.054 and 0.091 for trend, respectively). The researchers averaged 30-day rates of any hospital revisit and 30-day mortality over the study period and found that they were similar after discharge from ED treatment-and-discharge stays and observation stays (19.9% vs. 20.1% and 1.8% vs. 1.8%, respectively). For any hospital revisit and for death, rates were highest after discharge from inpatient stays (21.8% and 5.2%, respectively).

The authors noted that their study included only patients ages 65 years and older, did not report on such outcomes as symptoms or quality of life, and did not adjust data for case mix. However, they concluded that revisits to the hospital among Medicare beneficiaries, including those resulting in inpatient admissions, are common during the first month after observation stays and appear to be increasing in this group over time. They said that their results support further consideration of integrated care strategies for patients after observation stays, similar to those used following ED and inpatient stays, and called for further research in this area, as well as additional research on observation stays as a focus of quality improvement. “As observation stays become increasingly used as an alternative to short stay hospitalizations, further awareness of the quality of care and results associated with observation services may identify opportunities to improve outcomes for patients,” the authors wrote.

1 in 5 inpatient antibiotic regimens results in adverse drug event

About 20% of inpatients receiving antibiotics had an associated adverse drug event during hospitalization or shortly after discharge, according to a recent study.

The retrospective cohort study included 1,488 patients who received at least 24 hours of parenteral or oral antibiotic therapy while hospitalized at a single academic medical center. Their medical records were examined for antibiotic-associated adverse events: gastrointestinal, dermatologic, musculoskeletal, hematologic, hepatobiliary, renal, cardiac, and neurologic events within 30 days of antibiotic initiation and Clostridium difficile or incident multidrug-resistant organism infection within 90 days. Results were published by JAMA Internal Medicine on June 12 and appeared in the September print issue.

The study found that 20% of the patients experienced at least one antibiotic-associated adverse event during follow-up. Researchers also looked at whether the prescribed antibiotics were clinically indicated. They found that 19% of them were not (most commonly due to treatment of asymptomatic bacteriuria or noninfectious lower respiratory tract conditions), and that 20% of those non-indicated antibiotics were associated with an adverse event. Every additional 10 days of antibiotic therapy conferred a 3% increased risk of an adverse drug event, the researchers calculated. Gastrointestinal adverse events were most common, followed by renal and hematologic ones.

The study supports previous research about the frequency of adverse events with specific antibiotics, but produces “a more global picture of the overall incidence,” the study authors said. They noted that the study was limited by data from only a single center and that it should be replicated in other institutions and patient populations. The results may also represent an underestimate of antibiotic adverse events, given the hospital's existing stewardship efforts and the possibility that patients sought care outside the network included in the study data, the authors said. The study's results should be used by clinicians deciding whether to initiate or discontinue antibiotics and to support the importance of antibiotic stewardship, they added.

Another recent study looked at physician support of financial penalties for antibiotic misuse. More than 300 ACP members took an online survey and received one of four questions asking their positions on the issue and giving reasons for such penalties. When the question described harms to patients, including increased costs and iatrogenic infections, as the reason, 41% of the physicians supported penalties. When the harms described were increased costs to hospitals and insurers, 36% supported penalties. When societal harms, such as increased bacterial resistance to antibiotics and diversion of limited health care resources, were described, 23% supported penalties. In a control version that provided no information, 25% supported penalties. The results were published online June 20 by Annals of Internal Medicine and appeared in the Aug. 1 print issue.