Can Simple Laboratory Tests Predict Death as Well as Tried and True Clinical Data?

To better identify high-risk patients who need intensive therapy and low-risk patients who do not, laboratories are re-evaluating the role of simple, common laboratory tests. The complete blood count (CBC), electrolyte panel (EP) and creatinine are promising in this regard because they provide information on many different pathologies including anemia, bleeding, infection, cancer, kidney disease, and disturbances in water balance and acid base status.

In the abstract that I presented at the 2015 AACC meeting, my team examined the associations of these laboratory test results with mortality among patients undergoing coronary catheterization. The last laboratory test panel performed prior to catheterization was extracted from our city-wide laboratory information system and merged to outcome data from a clinical registry.

What we found was striking. A select group of red and white blood cell parameters, sodium, chloride and creatinine were strongly predictive of death, with a median of 2 years all-cause mortality after catheterization. When we collapsed test results into a 36 unit ‘lab risk score’, patients in the top 20% of score values had a 45-times higher risk of dying than those in the bottom 20%. The score also performed as well as a long list of clinical conditions – which interestingly did not improve risk prediction when used in addition to the lab risk score.

If laboratory data works as well as clinical data in predicting death, what is the value in scrutinizing simple lab test results if clinical data are available?

The answer is quality.

Pre-analytical processes aside, laboratory tests are performed and verified largely without human error. Quality control rules automatically monitor the quality of ongoing testing, as do proficiency challenges. Furthermore, lab results are stored in lab information systems which can be used to calculate risk scores to provide decision support on whether to treat or discharge.

The overall goal of our research is to both develop and validate a mortality risk score for coronary catheterization patients, and to determine the impact of its implementation on mortality rates.