Novel Risk Model Predicts Death 6 Months After MI in Elderly

Batya Swift Yasgur, MA, LSW

December 19, 2019

A novel risk calculation model that includes information about functional impairments and frailty shows promise in flagging 6-month mortality risk among elderly patients hospitalized with acute myocardial infarction (AMI).

"Our hope is that the model can be used at the time of discharge from an MI hospitalization in older patients to really understand mortality risk," senior author Sarwat I. Chaudhry, MD, associate professor of medicine, Yale University School of Medicine, New Haven, Connecticut, told | Medscape Cardiology.

The population is aging, and professional societies have endorsed risk prediction for the care of patients with AMI. Yet none of the risk models available for use for patients with MI have been developed or validated for older patients, she noted.

"What we felt was a potentially important omission from these earlier models is our growing sense that functional impairments may be important, and none of the previous risk models consider these things, which is what inspired us," Chaudhry recounted.

The investigators "wanted to find out what would happen if we took these functional impairments and other conditions of aging and lumped them all into a single heading called functional impairments. Would they rise to the top and stick in a final model after a very rigorous selection process?" she said.

The study was published online December 9 in the Annals of Internal Medicine.

Good Calibration, Discrimination

To investigate the question, the investigators identified 3006 participants (mean age, 81.5 years) in the prospective SILVER-AMI study of patients aged 75 years or older who were hospitalized with AMI at 94 US hospitals.

The team started out by examining 72 clinical variables associated with post-AMI mortality. Applying a rigorous statistical process, they narrowed down the list to a 15-variable model consisting of 11 standard risk factors (age, length of hospital stay, sleep apnea, peripheral artery disease, hemoglobin, heart rate, estimated glomerular filtration rate [eGFR], and revascularization on admission) and four additional factors — hearing impairment, mobility impairment, weight loss, and lower patient-reported health status — not included in previous risk models.

In the multivariable model, the strongest factors associated with 6-month mortality were receipt of in-hospital coronary artery bypass graft, which was protective (odds ratio [OR], 0.15); and age (OR, 1.06), length of stay (OR, 1.05), health status per level of worsening (OR, 1.34), and unintentional weight loss (OR, 1.69), which were associated with increased risk.

The final model had "good" discrimination in internal derivation and validation cohorts (area under the curve [AUC], 0.82 and 0.84, respectively) and was "well calibrated" (Hosmer–Lemeshow P > .05).

The investigators previously used this methodology to create an eight-variable risk prediction model for 30-day hospital readmission in older adults who have experienced AMI. As reported, impaired functional mobility was the strongest predictor of 30-day readmission; a Timed Up and Go (TUG) of between 15 and 25 seconds was linked to a 46% increase in the odds of readmission, and a TUG of 25 seconds or longer was linked to an 86% increase.

Reflection of Resilience

For the present analysis, the researchers compared the predictive performance of the SILVER-AMI score to that of the currently used Global Registry of Acute Coronary Events (GRACE) mortality risk score.

They found that discrimination was improved with the SILVER-AMI model compared with the GRACE model (AUC, 0.84 vs 0.76; P < .001).

Adding functional impairments significantly improved the predictive ability of the model, as evidenced by category-free net reclassification improvement indices of 0.21 for hearing impairment (P = .008) and 0.26 for mobility impairment, determined on the basis of TUG (P < .001)

"Looking at how quickly someone can walk, which might reflect information about the person's overall level of frailty, provides information about overall health and overall resilience, which is the information from an older person that is particularly powerful in suggesting level of risk and informing goals of care and shared decision making," Chaudhry commented.

Applying the Model

Chaudhry and colleagues illustrated the applicability of their model with respect to two hypothetical patients with identical clinical risk factors but different functional impairments/conditions of aging.

Both patients are 85-year-old men who undergo percutaneous coronary intervention. Admission heart rate is 90 beats/min, eGFR is 60 mL/min/1.73 m2, hemoglobin level is 120 g/L, ejection fraction is mildly decreased (45%), length of stay is 5 days, and there is no peripheral artery disease and no sleep apnea.

Their 6-month mortality risk, however, is vastly different when functional impairments/conditions of aging are accounted for using the SILVER-AMI model.

Patient A Patient B
Functional impairments/conditions of aging
  • TUG: ≤15 seconds

  • Hearing impairment: yes (a little)

  • Self-reported health status: good

  • Unintentional weight loss >10 lb: no

Functional impairments/conditions of aging
  • TUG: >25 seconds

  • Hearing impairment: yes (a lot)

  • Self-reported health status: fair

  • Unintentional weight loss >10 lb: yes

6-month mortality risk: 5% 6-month mortality risk: 22%


Multidimensional Assessment

Commenting on the study for | Medscape Cardiology, Michael W. Rich, MD, professor of medicine, Washington University School of Medicine, St. Louis, Missouri, said it "clearly demonstrates the incremental value of a multidimensional functional assessment for evaluating prognosis in older patients hospitalized with AMI."

The study has important take-home messages, he said.

"For older patients hospitalized with AMI ― or any condition ― it is critically important to consider nontraditional risk factors when assessing prognosis and designing care."

An accompanying editorial states, "Routine use of risk prediction models is valuable only if they actually enhance clinical decision making in a way that improves outcomes for older adults at the highest risk." Therefore, they "should be tested to determine whether they result in clinical improvement."

Chaudhry added, "If you see someone has a high risk of mortality due to a functional impairment like mobility, perhaps it is an impairment [on which] you want to focus more intervention, such as cardiac rehabilitation or PT [physical therapy]."

Patients at lower risk may "need less intensive interventions, while patients at higher risk may need more intensive medical therapy," she added.

The authors note that among other limitations, the prediction model was not externally validated, the mortality rate in SILVER-AMI was lower than in other published studies, and that the model requires information on functional impairments not typically assessed as part of routine inpatient care.

The SILVER-AMI risk predictor is available online at, Chaudhry noted.

The study was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health and by grant funding to individual investigators from the National institute on Aging. Chaudhry has received personal fees from the CVS Caremark Clinical Program for Connecticut. Coauthors' disclosures are listed in the original article. The editorialists' disclosures are available here. Rich has disclosed no relevant financial relationships.

Ann Intern Med. Published online December 9, 2019. Abstract, Editorial

For more Medscape Cardiology news, join us on Facebook and Twitter.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.