Recent Research

Gait speed and cardiac surgery outcomes, bloodstream infection surveillance, and more.


Gait speed preoperatively predicts cardiac surgery outcomes

Elderly patients with a slow gait speed have a threefold greater risk of experiencing a major complication or death following cardiac surgery than those who walk at a normal pace, and they are twice as likely to be discharged to another health care facility or have a prolonged postoperative stay.

Adding an assessment of the time it takes a person to walk five meters at a comfortable pace to existing cardiac surgery risk models appears to improve the predictive value of these models, helping clinicians identify vulnerable patients who might have been missed by using conventional measures such as ejection fraction alone, according to a study published in the Nov. 9, 2010 Journal of the American College of Cardiology.

Of the 131 patients included in the study (mean age, 75.8 ± 4.4 years), 60 (46%) had slow gait speed, defined as taking more than six seconds to walk five meters, using a cane or walker if needed. Treating physicians were blinded to the gait speed test results so as not to influence their decision to proceed with surgery or determine postoperative management.

Slow walkers had a less favorable profile overall, with higher death rates, a longer time to recovery and discharge from the hospital, and more need for rehabilitation facilities. Slow walkers were more often female (43% of women vs. 25% of men; P=0.03) and diabetic (50% vs. 28%, P=0.01).

Two different logistical regression models showed that slow gait speed improved the performance of the models in predicting morbidity and mortality. Independent predictors of discharge to a health care facility were slow gait speed (adjusted odds ratio [OR], 3.19; 95% CI, 1.40 to 8.41) and age 80 years and older (adjusted OR, 3.19; 95% CI, 1.19 to 8.60). Age 80 years and older was also an independent predictor of prolonged postoperative length of stay (adjusted OR, 2.95; 95% CI, 1.15 to 7.59). Slow gait speed showed a trend, but not a significant association toward longer length of stay (adjusted OR, 2.32; 95% CI, 0.95 to 5.67).

Slow walking appeared to be more predictive in female patients, with an adjusted odds ratio for mortality or major morbidity of 8.62 (95% CI, 1.46 to 51.00) in women and 1.65 (95% CI, 0.50 to 5.43) in men (P=0.18).

While walking speed should not be used in and of itself, measurement of gait speed does add incremental improvement to existing risk models, the authors concluded. An editorialist wrote, “Their new screening tool, consisting of an observer, a stopwatch and a well-lit hallway, is reproducible and I believe will be extraordinarily cost-effective.”

Sepsis often followed by decline

Hospitalization for severe sepsis was associated with increasing cognitive impairment and functional disability, according to a study of Medicare patients.

The prospective cohort study used data on hospitalizations and cognitive and functional assessments from the Health and Retirement Study, a survey of U.S. residents between 1998 and 2006. The study population included 516 patients hospitalized for severe sepsis and 4,517 patients who survived a nonsepsis hospitalization. Results were published in the Oct. 27, 2010 Journal of the American Medical Association.

After an episode of severe sepsis, patients' risk of having moderate to severe cognitive impairment increased more than threefold (odds ratio, 3.34, 95% CI, 1.53 to 7.25). The postsepsis patients also showed an increase in their number of functional limitations, especially if they had no or few limitations to start (mean of 1.57 new limitations in patients with no prior problems [95% CI, 0.99 to 2.15] and 1.50 new limitations for those with existing mild to moderate limitations [95% CI, 0.87 to 2.12]). These declines persisted for at least eight years.

The study also compared the sepsis patients to the cohort of patients that had been hospitalized for any other condition. Nonsepsis hospitalizations were not associated with developing moderate to severe cognitive impairment, and while the nonsepsis patients did develop some new limitations, they were significantly fewer than those found in the sepsis patients (P≤0.001 for the difference between sepsis and nonsepsis patients).

The cognitive and functional declines seen in these sepsis survivors represent a substantial, underrecognized public health problem, the study authors concluded. Extrapolating from the data, the authors estimated that in the U.S., 20,000 new cases of moderate to severe cognitive impairment may be attributable to sepsis annually. The complications may be partially preventable by greater efforts at sepsis prevention and improvements in sepsis care such as sedation management and early rehabilitation, the authors suggested. The results of the study may also be useful to physicians who are assessing care options and predicting outcomes for sepsis patients, an accompanying editorial said.

Quality of bloodstream infection surveillance varies widely

Bloodstream infection surveillance methods vary significantly across hospitals, which may affect reported infection rates and the ability to compare hospitals on this measure, a study found.

In a retrospective cohort study, researchers examined 20 ICUs in four medical centers between 2004 and 2007. Unit-specific central-line associated bloodstream infection (BSI) rates were calculated for 12-month periods. The data comprised 41 twelve-month unit periods representing 241,518 patient-days and 165,963 central-line days. Registered nurse infection control practitioners (“infection preventionists”) were blinded to the study, and performed routine prospective surveillance using definitions from the Centers for Disease Control and Prevention (CDC). Researchers then retrospectively applied a computer algorithm reference standard using criteria that approximated the CDC definitions. Results were published in the Nov. 10, 2010 Journal of the American Medical Association.

The median BSI rate as measured by infection control practitioners was 3.3 infections per 1,000 central line-days, compared to 9.0 infections per 1,000 central line-days as determined by the computer algorithm. The overall correlation between computer algorithm and practitioner rates was weak. There was significant heterogeneity among medical centers in the relationship between the algorithm and expected practitioner rates (P<0.001). The medical center with the lowest rate by routine surveillance (2.4 infections per 1,000 central-line days) had the highest rate by computer algorithm (12.6 infections per 1,000 central-line days).

Public reporting of hospital-specific infection rates is touted as a means to improve patient safety, with central line-associated BSI rates considered a key measure since these infections are common, costly, preventable and lead to poor outcomes, the authors noted. Interhospital comparisons are only valid, however, if surveillance methods are uniform and reliable; this study found strong evidence that such uniformity is weak. “Inconsistent surveillance practice can have a significant effect on the relative ranking of hospitals, which threatens the validity of the metric used by both funding agencies and the public to compare hospitals,” they wrote.

Customized computer program and posters reduced inpatient falls

A computerized fall prevention program successfully reduced falls among elderly inpatients, according to a study.

A trial was conducted over six months in four urban U.S. acute care hospitals and included about 10,000 patients in units. Half of the units provided usual care, including a fall risk assessment, and the other half implemented the intervention. On the intervention units, a clinician assessed a patient's fall risk on a computer through the project's fall prevention tool kit. Based on the results of the assessment, the tool kit automatically printed a customized poster to hang over the patient's bed, customized patient and family education materials, and a customized fall prevention plan for clinicians to implement. Control group patients received a generic warning poster if they were found to be at risk for falls and educational materials as needed.

During the six-month study, the intervention units had 67 patients fall compared to 87 on the control units (P=0.02). After adjustment for site, the fall rates in the intervention units were below those in the control units and the Massachusetts state average (3.15 per 1,000 patient days, 4.18 per 1,000 patient days [P=0.04], and 3.99 per 1,000 person days, respectively). Based on the results, study authors calculated that the intervention could prevent one fall per 862 patient-days, or if implemented in the study sites for a full year, one fall every four days, 7.5 falls each month, and about 90 falls in a year.

The intervention was likely effective because it was tailored to address patient-specific determinants of risk and was implemented within existing workflows. In addition, the intervention standardized communication about fall risk and made a prevention plan available at the bedside. The authors noted that a number of the alerts involved in the system related to patients' use of and ambulation to the toilet.

The fall prevention program was not found to be effective in patients under 65; fall rates didn't differ significantly between intervention and control groups for these younger patients. Future research should assess whether the lack of effect is a result of clinicians failing to perform recommended interventions for younger patients, or if different interventions are needed for younger patients, the researchers said. The study was published in the Nov. 3, 2010 Journal of the American Medical Association.

VTE recurrence more likely with nonsurgical transient risk factors

Risk for recurrent venous thromboembolism (VTE) is higher when the index event is provoked by a nonsurgical versus a surgical risk factor, but not high enough to warrant extended prophylaxis, according to a study.

It is usually considered safe to stop VTE prophylaxis three months after an index event triggered by a transient risk factor, such as pregnancy or surgery, but recurrence risk in this subgroup may vary by risk factor type. Researchers did a systematic review to determine rates of VTE recurrence after stopping anticoagulant therapy in patients whose VTE was provoked by any transient risk factor, surgery, or a nonsurgical risk factor. They also compared recurrence rates in patients with a first VTE related to a transient risk factor versus an unprovoked index VTE.

The researchers identified 15 prospective cohort and randomized trials for analysis, involving patients who had symptomatic VTE related to a transient risk factor and were treated for at least three months. Rates of recurrent VTE were calculated at 0- to 12-month and 12- to 24-month intervals after treatment was stopped. The study results appeared in the Oct. 25, 2010 Archives of Internal Medicine.

Twenty-four months after withdrawal of treatment, recurrence rates were 3.3% per patient-year for those with any transient risk factor, 0.7% per patient-year for those with a surgical transient risk factor and 4.2% per patient-year for those with a nonsurgical transient risk factor. After unprovoked VTE, the recurrence rate was 7.4% per patient-year. At 24 months, the authors calculated rate ratios of 3.0 for nonsurgical compared with surgical risk factors and 1.8 for unprovoked VTE compared with a nonsurgical risk factor.

Study limitations included use of differing definitions of “provoked” and “unprovoked” VTE and lack of continuous enrollment of patients with provoked VTE in the included trials. The authors concluded that risk for recurrence is highest for unprovoked VTE and that among those with transient risk factors, VTE provoked by a nonsurgical event confers a higher recurrence risk than VTE provoked by surgery. Still, the rate of recurrence in those with a nonsurgical transient risk factor is low enough to justify withdrawal of anticoagulant treatment after three months, per current recommendations, they wrote.

Expert consensus says PPIs, antiplatelets can be used together

Proton-pump inhibitors (PPIs) and antiplatelet drugs can be used together in cardiovascular disease patients who are at high risk of upper gastrointestinal (GI) bleeds, according to an expert consensus document drafted by three medical societies.

New recommendations include the following:

  • PPIs are recommended for patients with a history of upper GI bleeding, or for those with multiple risk factors such as a history of peptic ulcers, advanced age, Helicobacter pylori infection, or the use of antiplatelets, steroids or nonsteroidal anti-inflammatory drugs.
  • PPIs are not recommended to reduce upper GI bleeding in low-risk patients, and those who are less likely to benefit from prophylaxis.
  • Future studies are required to assess the impact of PPIs and antiplatelets among high-risk cardiac patients who poorly metabolize antiplatelet drugs.

The document was released in November 2010 by the American College of Cardiology (ACC), the American College of Gastroenterology (ACG), and the American Heart Association (AHA) in their respective journals and on their websites.

The three organizations recommended concomitant PPI and antiplatelet use in a 2008 document, but subsequent research suggested PPIs might lessen the efficacy of antiplatelet drugs. However, the consensus group noted, much of the published data used platelet function tests as surrogate markers of cardiovascular risk, and the clinical significance of laboratory test results has not been substantiated in large patient studies using clinically relevant endpoints, like heart attacks or strokes, a press release said.

COGENT, the only published, randomized controlled trial specifically designed to assess the clinical occurrence of GI bleeding, heart attacks and strokes associated with the prescription of clopidogrel alone versus clopidogrel with a PPI found no significant difference in cardiac events, the release said. This study, which included 3,761 patients with cardiovascular disease, demonstrated a 56% decrease in GI bleeding.

Prevention plan reduced ICU VAP

A multipronged prevention program lowered ventilator-associated pneumonia (VAP) rates compared to usual care in the ICU, but didn't eliminate the illness as researchers had hoped.

In a single-hospital study, the researchers compared VAP rates during a 45-month baseline period and a 30-month intervention period in 1,765 French patients who were mechanically ventilated for more than 48 hours. The intervention (prevention program) had previously been shown to increase compliance in eight areas: hand hygiene, gown-and-glove rules, keeping patients semirecumbent, maintaining an endotracheal tube cuff pressure >20 cm H2O, using an orogastric tube instead of a nasogastric tube, avoiding gastric overdistension, decontaminating the patient’ mouth with 0.12% chlorhexidine at least four times per day, and not using nonessential tracheal suction. VAP incidence rates were defined as total VAP episodes per 1,000 ventilator days and as first VAP episodes per 1,000 ventilator days at risk (until VAP diagnosis or extubation). Results appeared in the Nov. 15, 2010 Clinical Infectious Diseases.

Baseline and intervention VAP rates were 22.6 and 13.1 total episodes over total mechanical ventilation duration per 1,000 ventilator-days, respectively (P<0.001). Baseline and intervention rates were 26.1 and 14.9 first VAP episodes over mechanical ventilation duration at VAP or hospital discharge per 1,000 procedure-days, respectively (P<0.001). In both segmented regression analysis and a Cox proportional hazards model, VAP rates decreased by 43%, and remained significant after adjusting for confounders (Cox adjusted hazard ratio, 0.58; 95% CI, 0.46 to 0.72); P<0.001). The segmented regression analysis showed a trend toward increasing VAP rates over time before the intervention, and decreasing rates after the intervention. Daily VAP hazard rates on ventilation days 5, 10 and 15 were 2.6%, 3.5% and 3.4% during baseline, and 1.4%, 2.3% and 2% during the intervention.

The VAP rates at baseline were higher than are usually reported in multicenter surveillance networks, which may relate to France's more liberal defining criteria, the authors said. This high baseline rate may have increased the probability that the intervention would be successful, they noted. Still, VAP rates remained high even after the intervention, which “strongly suggest[s]” that completely eliminating VAP from the ICU may be “unreasonable” they concluded. Instead, hospital staff can use the measures in the prevention program to lower VAP incidence, without expecting complete elimination, they said.

Set of checklists helped reduce surgical complications, deaths

A comprehensive set of 11 surgical-safety checklists helped reduce complications and mortality at high-performing hospitals, a study found.

Researchers implemented the checklists at six Dutch hospitals with high standards of care. The checklists, which contained nearly 100 items, addressed issues like availability of imaging information, operative site verification and discharge instructions and covered the preoperative, intraoperative and postoperative periods. They compared outcomes of 3,760 patients observed before checklist implementation with 3,820 observed after implementation. Data were collected from a control group of five hospitals, as well. The article was published in the Nov. 11, 2010 New England Journal of Medicine.

The total number of complications per 100 patients fell from 27.3 pre-checklist to 16.7 post-checklist (absolute risk reduction, 10.6; 95% CI, 8.7 to 12.4). The proportion of patients with one or more complications decreased from 15.4% to 10.6% (P<0.001). In-hospital mortality decreased from 1.5% to 0.8% (absolute risk reduction, 0.7; 95% CI, 0.2 to 1.2). In the control hospitals, outcomes didn't change. Implementing these checklists requires much time and effort, given their comprehensive nature and the fact that they demand participation from many kinds of care providers, the authors noted. The study makes a “particularly persuasive case” for the effectiveness of checklists, because the intervention hospitals had already been tracking surgical outcomes, thus the results can't be attributed to performance feedback effects, an editorialist wrote. An important question to consider is whether improvements from checklists can be maintained in the long term, he noted.

Commercial drivers should stay off the road for a year after stroke

Patients who have had a transient ischemic attack (TIA) or stroke should not drive a commercial vehicle for at least the next year, determined a panel of experts advising the U.S. Department of Transportation.

The three-member panel conducted a systematic review on stroke and TIA survivors' risk for recurrent attacks and motor vehicle crashes. They found strong evidence that people who have had a TIA are at increased risk for stroke. Specifically, their pooled analysis found a 65-fold increased risk for stroke one month after TIA, which decreased to 16-fold at six months, and six-fold at one year. Less evidence was available regarding the motor vehicle crash risk of stroke survivors. Because the studies on this topic were small and of low-to-moderate quality, the experts concluded that there is “minimally acceptable” evidence that drivers who had a stroke are at increased, but unspecified, risk for a crash.

There is moderate evidence to suggest that neuropsychological testing can predict poststroke patients' driving performance, the experts concluded. In particular, the Figure of Rey, Motor-Free Visual Perception Test, and components of the Stroke Driver Screening Assessment were found to be significant outcome predictors in multiple studies.

Based on their review of the evidence, the panel concluded that patients who have suffered a TIA or minor stroke should not drive a commercial vehicle for at least one year, even if they are taking medication to prevent a stroke. After a year, individuals who want to drive commercially should undergo a thorough physical and mental evaluation by a neurologist or other physician knowledgeable on the topic.

The evaluation process should include an on-road driving evaluation, because it remains the gold standard for certification, the experts said. If the drivers pass the examination, they should be recertified on an annual basis and any recurrence of stroke or TIA will result in permanent disqualification. The panel's conclusions were published in the Dec. 1, 2010 Stroke.

Rule predicts imaging need with febrile UTI

A clinical prediction rule can help physicians determine when radiologic imaging is warranted in adults with febrile urinary tract infection (UTI), a study found.

Researchers in the Netherlands performed a prospective, observational study in consecutive adults who presented to eight emergency departments (EDs) with febrile UTI between January 2004 and November 2008. Ultrasound and computed tomography results were classified as normal or as indicating an urgent urological disorder (i.e., pyonephrosis or abscess), nonurgent urologic disorder, or incidental urological findings. Radiologic findings were considered clinically relevant if they indicated any urological disorder. Five EDs were used for the derivation cohort, while three were used for the validation cohort. Patients were followed for three months. The study results appeared in the Dec. 1 Clinical Infectious Diseases.

Of the 346 patients in the derivation cohort, 245 (71%) received radiologic imaging. The authors determined that history of urolithiasis, a urine pH of 7.0 or greater, and/or renal insufficiency (estimated glomerular filtration rate, ≥40 mL/min/1.73 m2) predicted both clinically relevant radiologic findings and urgent urological disorders (negative predictive values, 93% and 99%; positive predictive values, 24% and 10%). One point was assigned for each of these variables, with a possible score of 0 to 3. In the validation cohort, which included 131 evaluable patients, the negative predictive values for clinically relevant radiologic findings and urgent urological disorders were 89% and 100%, respectively, while the positive predictive values were 20% and 11%. In total, 339 of 490 patients (69%) received radiologic imaging. If imaging had been performed only in the 203 patients with a score of 1 or higher on the clinical prediction rule, implementing the prediction rule might potentially lead to 40% relative reduction or 28% absolute reduction in radiologic imaging studies.

The study authors called for further research examining the modification and validation of their rule in other settings and bigger cohorts, and prospective studies examining the rule's cost-effectiveness. For now, they advocate implementing the prediction rule “because it predicts urgent urological complications at presentation rather than after 3 days of treatment.”