Postop deaths, sepsis projects

Summaries from ACP Hospitalist Weekly.

Nearly all deaths related to noncardiac procedures occur after surgery, study finds

The vast majority of deaths related to noncardiac surgery occur after the procedure, with nearly half being associated with three complications, a prospective cohort study found.

Image by Getty Images
Image by Getty Images

The international Vascular Events in Noncardiac Surgery Patients Cohort Evaluation (VISION) study included 40,004 patients ages 45 years and older (49.7% women; mean age, 63.9 years) who had inpatient noncardiac surgery at 28 centers in 14 countries. Researchers monitored patients for complications until 30 days after surgery and assessed the association between complications and 30-day mortality. They also determined the attributable fraction for each complication to estimate the potential proportion of deaths attributable to it. Results were published online on July 29, 2019, by the Canadian Medical Association Journal.

A total of 715 (1.8%) patients died within 30 days of surgery, and 99.3% of deaths occurred after the procedure. Five (0.7%) patients died in the operating room, 500 (69.9%) died after surgery during the index admission, and 210 (29.4%) died after discharge. The median time to death was 11 days. Overall, 44.9% of deaths were associated with three complications: major bleeding (6,238 [15.6%] patients; adjusted hazard ratio [HR], 2.6; 95% CI, 2.2 to 3.1), myocardial injury after noncardiac surgery (5,191 [13.0%] patients; adjusted HR, 2.2; 95% CI, 1.9 to 2.6), and sepsis (1,783 [4.5%] patients; adjusted HR, 5.6; 95% CI, 4.6 to 6.8). These three complications had the largest attributable fractions of deaths (17.0%, 15.9%, and 12.0%, respectively) and occurred a median of 0, 1, and 6 days, respectively, after the procedure.

Among the study's limitations is the fact that some outcomes were not adjudicated, including major bleeding, sepsis, infection without sepsis, congestive heart failure, and acute kidney injury with dialysis, the study authors noted. They concluded that focusing on the prevention, early identification, and management of major bleeding, myocardial injury after noncardiac surgery, and sepsis holds promise for reducing perioperative deaths. “The median time to these events provides insight about when monitoring for each complication is likely to have the greatest effect,” they wrote.

However, bias in observational studies is almost unavoidable, and the findings must be interpreted carefully before they are applied to clinical and policy decisions, a commentary noted. “I salute the achievement of the VISION study investigators but advise caution in applying the relationships of complications with 30-day mortality and the attributable fractions,” the editorialist wrote.

Sepsis improvement efforts have reduced mortality, two studies find

Efforts in recent years to improve sepsis care have been successful, according to two recent studies.

The first was a retrospective cohort study of more than 1 million adult patients hospitalized with sepsis in New York State or any of four control states (Florida, Maryland, Massachusetts, and New Jersey) from Jan. 1, 2011, to Sept. 30, 2015. The study was designed to analyze the effects of the New York State regulation that required hospitals develop, implement, and report on protocols to recognize and treat sepsis beginning in 2013. Results of the study were published in the July 16, 2019, JAMA.

Unadjusted 30-day in-hospital mortality was 26.3% in New York and 22.0% in the control states before the regulations, and it decreased to 22.0% in New York and 19.1% in control states in the postregulation period. After adjustment, the decrease in New York State was significantly greater than that in the control states (P=0.02 for the joint test of the comparative interrupted time-series estimates). The study also compared how a number of secondary outcomes changed in New York versus the other states and found no significant change in rates of ICU admission, an increase in central venous catheter use, and decreases in Clostridium difficile infection rates and hospital length of stay.

The results support the effectiveness of the New York State sepsis regulation, the authors said. “At the same time, the observed patterns of sepsis mortality raise concerns about the generalizability of the findings outside of New York State. . . . the regulations may have served primarily to correct relatively poor-quality sepsis care in New York State,” they wrote. It's uncertain whether other states could see the such improvement from similar regulations, according to the authors.

An accompanying editorial said that the New York example “illustrates a rarely used but potentially effective public health intervention for reducing mortality and morbidity from sepsis” but also cautioned that mandated guidelines and protocols can have negative effects. The editorialist called for additional research into the outcomes and costs of such regulations, concluding that “any strategy aimed toward reducing sepsis-related morbidity and mortality must be based on convincing evidence before being mandated by governmental regulations.”

The other study was an observational study including 13,877 adults with suspected infection treated at a single medical center in Illinois between March 1, 2014, and July 31, 2017. This time period included 18 months before and after CMS instituted the SEP-1 performance measure. At that time, the hospital implemented a bundle of interventions to improve care for sepsis: a multidisciplinary sepsis committee with sepsis coordinator and data abstractor, an education campaign, electronic health record tools, and a modified early warning system. Results were published by Critical Care Medicine on July 15, 2019, and appeared in the October 2019 issue.

Risk of in-hospital death was significantly lower among patients with suspected infection in the postimplementation period (odds ratio, 0.70; 95% CI, 0.57 to 0.86) as was time to discharge. Both improvements were also seen in the subgroup of patients who met Sepsis-3 criteria. The program was also cost saving in patients with suspected infection (–$272,645.7; 95% CI, –$757,970.3 to –$79,667.7) and those in the Sepsis-3 group.

The study authors concluded that their effort to adhere to the SEP-1 measure led to improvements in care and costs for an even larger group of patients than was targeted by CMS and that the results might serve as a benchmark for other quality improvement efforts. “In all regards, our quality improvement program led to better adherence for early treatment of sepsis indicated by a reduction in time-to-first-antibiotic administration, a hallmark for improving mortality in sepsis care,” they said. Limitations include the risk of confounding and the uncertain generalizability of results from a single center.

Two-thirds of patients admitted for pneumonia received excess antibiotic therapy

While shorter antibiotic duration is safe and effective for general medicine patients admitted with pneumonia, excess treatment is common, especially at discharge, a retrospective cohort study found.

Researchers used data from 43 hospitals in the Michigan Hospital Medicine Safety Consortium to assess the rate of excess antibiotic treatment duration (excess days per 30-day period) in 6,481 adult general medicine patients (median age, 70.2 years; 51.2% women) hospitalized with community-acquired pneumonia (CAP; 73.2% of cohort) or health care–associated pneumonia (HCAP; 26.8% of cohort). Eligible patients had a discharge diagnostic code for pneumonia, symptoms and radiographs consistent with pneumonia, receipt of at least four days of antibiotic treatment, and receipt of antibiotics on hospital day one or two (to exclude hospital-acquired pneumonia). Study participants were discharged between January 2017 and April 2018.

Excess days of antibiotics were calculated by subtracting each patient's shortest effective treatment duration (as was expected based on time to clinical stability, pathogen, and pneumonia classification) from the actual duration. Patients with CAP were expected to have a treatment duration of at least five days (lengthier if time to clinical stability was longer), whereas patients with HCAP, Staphylococcus aureus, or a nonfermenting gram-negative bacillus were expected to have a treatment duration of at least seven days. At 30 days, researchers assessed patient outcomes through medical records and telephone follow-up. Results were published online on July 9, 2019, by Annals of Internal Medicine and appeared in the Aug. 6, 2019, issue.

Overall, 4,391 (67.8%) patients received excess antibiotic therapy, 71.8% (3,410 of 4,747) of those with CAP and 56.6% (981 of 1,734) of those with HCAP. Antibiotics prescribed at discharge accounted for 93.2% of excess days of therapy. While 99.6% of patients had an expected therapy duration of five or fewer days at discharge, five days was the most common discharge prescription. “This highlights an urgent and unmet need for ‘discharge stewardship,’ or coordinated interventions to improve antibiotic prescribing at discharge,” the study authors wrote.

Patients who were more likely to receive excess treatment included those who had respiratory cultures or nonculture diagnostic testing, had a longer stay, had received a high-risk antibiotic in the prior 90 days, had CAP, or did not have a total antibiotic treatment duration documented at discharge. Excess treatment was not associated with lower rates of adverse outcomes, including death, readmission, or ED visit. Furthermore, each excess day of antibiotic treatment was associated with a 5% increase in the odds of related adverse events reported by patients after discharge, most often diarrhea, gastrointestinal distress, and mucosal candidiasis.

Limitations of the study include its retrospective design and the fact that patient-reported adverse events are subject to recall bias and inaccurate reporting, the study authors noted. “Together, these findings provide evidence that reducing excessive antibiotic treatment durations, especially in patients with CAP via discharge stewardship, is safe and may improve patient care,” they wrote.

The study results add to a considerable body of evidence supporting the antibiotic mantra “shorter is better” for not only pneumonia but also acute exacerbation of chronic bronchitis and sinusitis, complicated urinary and intra-abdominal infections, gram-negative bacteremia, acute bacterial skin infections, osteomyelitis and septic arthritis, and even neutropenic fever, an accompanying editorial noted. More than 45 randomized controlled trials and two meta-analyses have found no difference in efficacy between shorter and traditional courses of antibiotic therapy for these diseases, the editorialists said.

“Change is scary, and medicine is a conservative profession. . . . Nevertheless, to live up to the expectations that our patients have for us and that we have for one another, we must overcome inertia and tradition and change practice when compelling evidence becomes available,” they wrote.

Penalties for hospital-acquired conditions didn't lead to improvements in safety

Hospitals that incurred Medicare penalties for hospital-acquired conditions (HACs) did not show significant improvements in their subsequent HAC rates, a recent study found.

The retrospective cohort study used regression discontinuity design to look at more than 15 million discharges of Medicare fee-for-service beneficiaries from 3,238 acute care hospitals between July 23, 2014, and Nov. 30, 2016, that were eligible for at least one HAC targeted by the penalties. It included 708 of the 724 hospitals penalized for HACs during fiscal year 2015. Results were published by The BMJ on July 3, 2019.

The penalized hospitals had 2.72 HACs per 1,000 episodes compared to 2.06 per 1,000 episodes in nonpenalized hospitals. Penalized hospitals were more likely to be large, to be teaching institutions, and to have a greater share of patients with low socioeconomic status. Rates of 30-day readmissions were 14.4% in penalized hospitals and 14.0% in nonpenalized ones, and 30-day mortality was 9.0% in both categories. Penalties were not associated with significant improvements in HACs (−0.16 per 1,000 episodes; 95% CI, −0.53 to 0.20), readmissions (−0.36 percentage point; 95% CI, −1.06 to 0.33), or mortality (−0.04 percentage point; 95% CI, −0.59 to 0.52).

Previous research had already shown that the HAC penalties particularly affect major teaching hospitals and facilities caring for disadvantaged patients, according to the study authors. “Our findings add to this literature by demonstrating that penalization does not appear to drive performance improvement in the program,” they said. Together these results suggest that the HAC penalty program could be decreasing equity while not improving quality, the authors noted.

They offered a number of potential solutions to this problem, including changing the HAC penalties to be gradated instead of all or nothing for the hospitals in the bottom quarter for performance and modifying penalties based on hospitals' share of disadvantaged patients. The authors also called for additional research to determine whether future measurement and penalty programs are succeeding in improving patient safety.

Index for preoperative cardiovascular evaluation may have strong predictive power

A recently developed index for preoperative cardiovascular evaluation of patients undergoing noncardiac surgery showed strong discriminatory power to stratify patients as low, intermediate, or high risk, a study reported.

To develop the Cardiovascular Risk Index (CVRI), researchers used 3,284 prospectively enrolled adult patients undergoing noncardiac surgery at a hospital in Beirut. Multivariate logistic regression analysis identified six easily acquired, predictive data elements: age 75 years or older, history of heart disease, symptoms of angina or dyspnea, hemoglobin level below 12 mg/dL, vascular surgery, and emergency surgery.

The validation cohort consisted of 1,167,414 patients registered in the American College of Surgeons National Surgical Quality Improvement Program database as having had noncardiac surgery between 2008 and 2012. The primary outcome measure was death, myocardial infarction, or stroke at 30 days after surgery. Results were published in the June 25, 2019, Journal of the American College of Cardiology.

The primary outcome occurred in 38 patients (1.2%) in the derivation cohort. Patients were categorized by how many of the six characteristics they met: zero, one, two, three, or more than three. The incidence of the primary outcome increased steadily across these CVRI groups in both the derivation (0%, 0.5%, 2.0%, 5.6%, and 15.7%, respectively; P<0.0001) and validation (0.3%, 1.6%, 5.6%, 11.0%, and 17.5%, respectively; P<0.0001) cohorts. The discriminatory power of the new CVRI was further confirmed by constructing a receiver-operating characteristic curve that had an area under the curve of 0.90 in the derivation cohort and 0.82 in the validation dataset.

This has important implications for the efficient triage and management of patients scheduled for noncardiac surgery, authors wrote. “Moreover, the new index is able to identify a large group of very-low-risk patients that, in general, might not require pre-operative cardiovascular evaluation or any special monitoring postoperatively.”

An editorial noted that the potential to identify and treat comorbid conditions may partly explain the performance of the CVRI, because its variables reflect heart disease symptoms of advanced or poorly compensated conditions, as well as anemia, a factor known to be independently associated with post-operative mortality. “Beyond its intended purpose for risk stratification, the CVRI may encourage constructive measures to be taken during pre-operative evaluation by highlighting conditions amenable to intervention.”

Procalcitonin may not be reliable enough to withhold or mandate antibiotics in community-acquired pneumonia

Serum procalcitonin levels are neither sufficiently sensitive nor specific to distinguish bacterial from viral community-acquired pneumonia (CAP), according to a systematic review and meta-analysis.

Researchers evaluated 12 studies with a total of 2,408 patients. Included studies specifically enrolled adults with CAP and stratified patients based on presumptive or proven etiologic diagnoses of viral, bacterial, or atypical infection. Results were published online on June 25, 2019, by Clinical Infectious Diseases.

Four studies included only ICU patients, and the remaining studies included all patients who were hospitalized for CAP, except for one that also included patients who had acute exacerbation of chronic obstructive pulmonary disease. Studies used different thresholds of serum procalcitonin levels to distinguish bacterial from nonbacterial CAP, with 0.5 µg/L being the most relevant cutoff point.

In eight studies that used the 0.5 µg/L cutoff, the pooled sensitivity and specificity estimates were 0.55 (95% CI, 0.37 to 0.71) and 0.76 (95% CI, 0.62 to 0.86), respectively. In general, studies that showed the greatest sensitivity showed the lowest specificity and vice-versa. The hierarchical summary receiver-operating characteristic curve had an area under the curve of 0.73 (95% CI, 0.69 to 0.76), a moderate result.

The results suggest that when managing patients with CAP, a serum procalcitonin level is “unlikely to provide information that will enable clinicians to immediately address whether the infection is bacterial and antibiotics need to be administered, or whether it is viral and antibiotics may be withheld,” the study authors wrote. “The sensitivity and specificity are both too low and variable for the results to be confidently used in the decision-making process.”

Limitations of the study include the potential for publication bias and the restrictive exclusion criteria, which may have led to the omission of valid studies, the authors said. In addition, most of the early procalcitonin-guided therapy trials took place in Europe, where the proportion of CAP caused by Streptococcus pneumoniae substantially exceeds that in the U.S., they noted.

Most Medicare readmissions were for diagnoses different from index hospitalization

Utilization of postdischarge services was strongly associated with preventing 30-day readmissions among Medicare fee-for-service beneficiaries, whether or not the services were for the same diagnosis as the index hospitalization, a recent study found.

Researchers used U.S. administrative claims data to assess the relationship between unplanned 30-day readmission and health services utilization after discharge for the same or a different diagnosis than the index hospitalization. Results were published online on June 21, 2019, by the Journal of General Internal Medicine and appeared in the September 2019 issue.

Image by Getty Images
Image by Getty Images

The study included 583,199 all-cause index hospitalizations of Medicare fee-for-service beneficiaries in 2014. Researchers also evaluated individual index hospitalizations for heart failure, myocardial infarction, and pneumonia. Postdischarge services utilization included care from institutional outpatient, office-based primary care, office-based specialist, office-based nonphysician, emergency department, home health care, and skilled nursing facility clinicians.

Among all-cause index hospitalizations, 11.7% resulted in 30-day readmissions, only 18.1% of which occurred for the same primary diagnosis as the index hospitalization. The majority of postdischarge health services were utilized for a primary diagnosis differing from index hospitalization. Compared with no visit, institutional outpatient visits for the same as or a different primary diagnosis from index hospitalization were similarly strongly associated with decreased 30-day readmission (odds ratios, 0.33 [95% CI, 0.31 to 0.34] and 0.36 [95% CI, 0.35 to 0.37], respectively). There were similar patterns for primary care physician, specialist, nonphysician practitioner, and home health care. In addition, index hospitalizations for heart failure, myocardial infarction, and pneumonia showed similar patterns to all-cause index hospitalizations.

Limitations of the study include the possibility of unmeasured confounders (e.g., selection of specific postdischarge services) and a lack of information on patients' socioeconomic status, the study authors noted. In addition, the results may not be generalizable to patients who are not Medicare fee-for-service beneficiaries, they said.

“Our study suggests that physicians should consider a holistic approach to post-discharge care because utilization of post-discharge services of various types and for various diagnoses contributes to reduced readmission,” the authors wrote. “Furthermore, medical practitioners should consider patients' financial and non-financial (e.g., travel time) constraints when planning post-discharge care to optimize access to the various health services required.”

Admission to an ICU associated with lower 30-day mortality in STEMI patients

Admission to an ICU was associated with lower mortality among patients who had ST-elevation myocardial infarction (STEMI) and marginal need for intensive care, a recent study found.

The retrospective cohort study included 109,375 Medicare beneficiaries (ages 65 years or older) admitted with STEMI to 1,727 acute care hospitals between January 2014 and October 2015. Researchers assessed the rate at which hospitals admitted such patients to the ICU; the hospitals in the top quarter of ICU admission rates admitted 85% or more of their STEMI patients to an ICU. Results were published by The BMJ on June 4, 2019.

Using an instrumental variable analysis, researchers identified patients whose proximity to a hospital in the top quarter of ICU admission rates appeared to determine whether they were admitted to an ICU. Among those patients, being admitted to an ICU was associated with lower 30-day mortality compared to being admitted to a general, telemetry, or intermediate care ward (absolute decrease, 6.1 percentage points; 95% CI, −11.9 to −0.3 percentage points). The study also looked at patients with non-ST-elevation MI (NSTEMI) as a comparison population, since routine ICU use has been shown not to improve their outcomes, and found that ICU admission for NSTEMI was not associated with reduced mortality (absolute increase, 1.3 percentage points; 95% CI, −0.9 to 3.4 percentage points).

“Contrary to the prespecified hypothesis, we found that ICU care may be underused for certain patients with STEMI,” the authors wrote. They noted that the results apply only to patients without obvious indications for or against ICU care and that the study did identify some population-level characteristics of such “marginal” patients, including age older than 85 years and no organ failure.

The results do not show what about the ICU was particularly beneficial to the patients, but the authors speculated that it could be enhanced nursing care, better management of noncardiac conditions, timely access to certain treatments, or more effective protocols. “Thus, the results of this study may have more to do with limitations of non-ICU care rather than the direct benefits of ICU care,” the authors wrote. Limitations of the study include the risk of residual confounding, the treatment of ICUs and coronary care units as equivalent, and the inclusion of only patients 65 years of age and older.

“Conventional wisdom in the US suggests that ICU care is generally overused and that efforts must be made to reduce the number of patients receiving ICU care. However, this study, in combination with others, indicates instead that ICU care may often be misdirected, with some patients experiencing underuse while others experience overuse,” the authors concluded.