Journal Wrap | Kim Eagle, MD, and the editors of ACC.org, present relevant articles taken from various journals.
Does Chronic Lung Disease Affect TAVR Outcomes?
Patients with moderate to severe chronic lung disease (CLD) have an increased risk of death up to 1 year after Transcatheter Aortic Valve Replacement (TAVR), according to a recent study published in The Annals of Thoracic Surgery. In patients with severe CLD, the risk of death is similar with either transaortic or transapical approaches.
For this study, clinical records for 11,656 patients undergoing TAVR from 2001 to 2014 in The Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy (TVT) Registry were linked to Medicare hospital claims. Overall, the median age of patients was 84 years and 51.7% were female. In addition, 3,225 (27.7%) of patients had moderate or severe CLD.
Results found that patients with severe CLD were younger, more likely to have symptoms of advanced heart failure, more likely to have a history of smoking, and more likely to be 02 dependent than those with mild or no CLD. Patients with severe CLD also spent significantly more time in the intensive care unit than patients with moderate and no or mild CLD, and they were significantly more likely to be discharged to an extended care or rehabilitation hospital. Additionally, severe CLD patients had a significantly higher unadjusted incidence of in-hospital mortality, however, no significant difference was found in the unadjusted incidence of stroke. Patients with moderate CLD did not experience increased odds of in-hospital death.
Up through 1 year, patients with moderate and severe CLD experienced an increased risk of mortality (but not an increased risk of stroke) compared to patients with no or mild CLD. In patients with severe CLD, home 02 was present in 47.6% of patients and pulmonary hypertension was present in 28.6% of patients. Both of these were associated with an increased adjusted risk of mortality up to 1 year. Home 02 was associated with a lower adjusted 1-year risk of stroke, but pulmonary hypertension was not associated with stroke risk.
The researchers also evaluated the subgroup of patients with non-transfemoral access in order to determine the association between the transaortic and transapical approaches and outcomes. No significant differences were noted in the length of intensive care, incidences of in-hospital death, in-hospital stroke, or discharge location between the approaches. Additionally, no significant difference was seen in adjusted 1-year mortality.
The authors conclude that “further research is necessary to understand strategies to mitigate risk associated with CLD and the long-term implications of these findings.”
Suri RM, Gulack BC, Brennan JM, et al. Ann Thoracic Surg. 2015;doi: 10.1016/j.athoracsur.2015.05.075. [Epub ahead of print]
Can a Flu Vaccine Protect the Heart?
Patients with influenza (flu) infections may be at a higher risk of acute myocardial infarction (AMI), but a flu vaccination may lower that risk, according to a recent study published in Heart.
The relationship between AMI and the flu has been known since the 1930s, after increased cardiovascular deaths were seen during the flu season. While many countries recommend flu vaccinations for patients at increased risk of complications, including those with cardiovascular disease, vaccine coverage remains suboptimal in this population.
Researchers analyzed 16 case-controlled studies—8 on flu vaccination, 10 on flu infection and AMI. They found that flu infection was significantly associated with AMI, with cases having double the risk of flu infection or respiratory tract infection compared with controls. The data also showed a 29% pooled vaccination effectiveness in preventing AMI, which is similar to the efficacy of other secondary prevention measures such a statins, antihypertensives and smoking cessation interventions.
“Given the high global burden of AMI, and ischemic heart disease being the leading cause of death and disability in the world, influenza vaccination could be added to other preventative strategies and confer additional population health benefits on AMI prevention,” the authors write. They add that vaccination is inexpensive, safe and effective. Given that the risk of AMI increases after 50 years of age, these findings also add to the evidence base that middle-aged adults should receive a flu vaccine. However, the authors note that the interpretation of vaccine effectiveness is complex and the vaccination may not be equally protective against AMI the entire year. Four of the six studies they examined on vaccinations were performed during the flu season. Additionally, the effectiveness of the annual flu vaccine varies depending on the circulating strain. The timing of the vaccine is also important. The vaccine must be administered prior to the AMI event for it to be a valid predictor of AMI risk.
The authors stress that physicians should be aware of the need to offer flu vaccinations to patients with cardiovascular disease, and that cardiologists should consider offering vaccination following an AMI, prior to hospital discharge or during cardiac rehabilitation or follow-up. Additionally, cost-effectiveness studies are needed to compare the flu vaccine and primary and secondary prevention for AMI in order to further inform preventative policy.
Barnes M, Heywood AE, Mahimbo A, et al. Heart. 2015;0:1-10.
Researchers Develop First Outcome Measures for EHRs
A study recently published in Medical Care maps out the first outcome measure to use with electronic health records (EHRs), measuring 30-day mortality after acute myocardial infarction (AMI).
Currently, the Centers for Medicare and Medicaid Services (CMS) publicly reported hospital outcome measures use claims data for risk adjustment, which many clinicians have expressed concerns with. With the support of CMS, Robert L. McNamara MD, MHS, FACC, and colleagues developed an eMeasure of hospital 30-day all-cause risk-standardized mortality for patients with AMI. Their goal was to create an outcome measure that would be suitable for national public reporting by using data elements that are routinely collected in current practice, captured in standards formats and feasibly retrieved from current EHR systems. The developed eMeasure was recently endorsed by the National Quality Forum (NQF).
Researchers used data from the ACC/American Heart Association Action Registry Get With the Guidelines, which captures information about AMI patients at 450 hospitals nationwide. They matched admissions for AMI in the registry that were discharged between Jan. 1 and Dec. 31, 2009, with admissions for AMI from CMS Medicare Part A discharged during the same time period. The researchers also derived a cohort for the validation in a similar fashion based on the claims between Jan. 1 and Dec. 31, 2010. They then calculated the hospital-specific risk-standardization mortality rate at the ratio of predicted number of death to expected number of deaths, multiplied by the nation adjusted mortality rate. Then, they calculated the correlations between the risk-standardization mortality rate from the final model and those from CMS’s publicly reported claims-based AMI mortality measure.
The final 2009 cohort consisted of 20,540 discharges from 280 hospitals. Among admissions eligible for matching in Action-GWTG, 75% were successfully matched in the CMS claims data. Among admissions eligible for matching in the CMS claims dataset, 53% were successfully matched with the Action-GWTG data. The final logistic regression model included the five variables determined to be eMeasure-feasible—age, heart rate, systolic blood pressure, troponin ratio and creatinine—and had a similar discriminative performance with an area under the receiver operative characteristic (ROC) curve of 0.78. The estimated between-hospital variance in the log-odds mortality was 0.025, which implies that the odds mortality for a high-mortality hospital were 1.37 times that of a low-mortality hospital. Both the 2009 development and 2010 validation models exhibited strong discrimination, calibration and fit, and predictive ability was similar across datasets.
In the 2009 development cohort, the hospital unadjusted 30-day hospital mortality rate ranged from 0% to 60% with a median of 10.5%. After adjusting for patient characteristics and clustering within hospitals, risk-standardization mortality rates at the hospital level were normally distributed, ranging from 9.6% to 13.1%, and the median risk-standardization mortality rate was 10.7%. The correlation coefficient between risk-standardization mortality rates from the final model and risk-standardization mortality rates from the claims-based AMI mortality measure based on 2009 data was 0.86, demonstrating an excellent correlation.
“This measure of 30-day risk-standardized AMI mortality rates for Medicare beneficiaries was developed to be feasible for implementation in current EHR systems and consistent with the current standard clinical practice,” the authors write. “We established novel criteria to assess feasibility of data elements for inclusion in an eMeasure.” They add that the final model contains five risk-adjustment variables available on or near patient presentation that performed very well and produced estimated comparable with the previous 27-variable CMS claims-based mortality measure for AMI. They believe that this eMeasure provides a blueprint for the development of future eMeasures for a wide variety of conditions. “However, current practice and current EHR systems have not optimized standardization of data collection, limiting the potential. Increasing standardization should allow assessment and incorporation of additional clinical elements not currently feasible for use in the EHR environment to create even more powerful outcome-based eMeasures. These functional eMeasures will allow real time collection of data with rapid feedback for clinicians and health administrators, facilitating quality improvement.”
McNamara and colleagues add that their methodology reflects accepted standards for publicly reported outcome measures. This includes a standard time frame that ensures hospital variation in length of stay does not affect performance and minimizes opportunities for misrepresentation. The 30-day timeframe may lead to better collaboration among hospitals and medical communities, leading to a reduction in mortality rates.
The ability for EHRs to accurately retrieve required data is critical to implementing eMeasures. The authors tested and confirmed that the majority of criteria and variables were accurately extracted from current EHRs. However, there were exceptions, including identifying patients who had been transferred from another facility among others, that highlight the need for considering performance measure requirements in future EHR development.
McNamara RL, Wang Y, Partovian C, et al. Med Care. 2015;53:818-26.
< Back to Listings