A point-based prediction model for cardiovascular risk in orthotopic liver transplantation

VanWagner, L. B. et al. (2017), A point-based prediction model for cardiovascular risk in orthotopic liver transplantation: The CAR-OLT score. Hepatology, 66: 1968–1979. doi:10.1002/hep.29329

BACKGROUND

Cardiovascular disease (CVD) complications are important causes of morbidity and mortality after orthotopic liver transplantation (OLT). Previous prediction algorithms are not made user friendly. Currently there is no pre-operative risk assessment tool to estimate risk of CVD events after OLT. A risk score was developed called the CArdiovascular Risk in Orthotopic Liver Transplantation (CAR-OLT) score.

METHODS

Adult, first liver or liver-kidney transplant for chronic liver disease/primary cancer at Northwestern Medicine (NWM), Chicago between 2002-2012 (n=1024, 88% n=904 alive 1 year post OLT were reviewed. Outcomes were major CVD complications within 1 year OLT (primary) or other CVD complications (secondary). Multivariate analysis was used. Race, sex, age were forced into the final model.

RESULTS

Mean age was 56.3, 65.9% were male, 78.6% were non-Hispanic white. Reasons for transplantation were Hepatitis C (33.5%), alcohol (23.5%), NASH (16.1%). The average MELD score was 20.9. Pre-existing CVD conditions were common. Overall characteristics were similar to nationally. There were 498 CVD complications among 329 (32.1%) patients within 1 year (majority within 30 days). Most common was heart failure, then atrial fibrillation, then stroke.

The CAR-OLT score generated included age, sex, race, working status, education, AF, respiratory failure on a ventilator, pulmonary hypertension, hepatocellular carcinoma, hypertension, diabetes mellitus, heart failure.

The majority of patients had a CAR-OLT score <40 points and a predicted probability of a one year CVD complication <40%. However, a significant proportion of patients (30.1%) were at very high risk (predicted risk >45%). The model demonstrated good discrimination with the number of predicted generally matched the number of observed complications indicating excellent calibration.

DISCUSSION

The paper makes comments on the ease of use and originality of the CAR-OLT score compared to previous scores. It also states that employment appears to be a surrogate for functional status.

Strengths: Size, rigorous assessment, generalisability, state of the art methods to evaluate performance of prediction model.

Weaknesses: Only CVD events at NWM hospital and in the top three codes which may underestimate incidence, inaccurate history taking may have occurred, laboratory values were not always available, single centre. The score needs validation in a multicentre trial.

Discussion in the journal club included that the results were clearly defined in an appropriate population. It has not been validated to different populations and it remains unclear whether the score would transfer to the UK population and practices. Statistical tests are appropriate, however some aspects e.g. area under operator curves are not described in detail. Calibration appears excellent. However, this occurs against the study population so this would appear inevitable. Whether it calibrates well to the wider population is unknown. It remains unclear whether the score would modify one’s decision-making as its simplicity is irrelevant given the extensive investigations and multi-disciplinary team decision making utilised for liver transplantation.

Summary by Dr Jim Parry

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s