The Liverpool causality assessment tool versus the Naranjo Scale in assessment of suspected adverse drug reaction - a retrospective chart review study

M. Rieder1, S. Um2, A. Abuzgaia1. 1Paediatics, Physiology & Pharmacology and Medicine, Western University, London, CANADA, 2Family Medicine, Western University, London, CANADA.

Objective

Evaluation of possible adverse drug events is difficult and there are several potential screening tools that can be used. We assessed the clinical value of the Liverpool causality assessment tool in comparison to the Naranjo scale for screening suspected adverse drug reaction (ADR) cases.

Method

We retrospectively reviewed patient charts who presented with suspected history of ADR, and score using the Liverpool causality assessment tool and the Naranjo scale and determine how well each tool correlates with laboratory and other investigations.

We reviewed charts from the Clinical Pharmacology Clinic at London Health Sciences Centre (LHSC) from 2008 to 2015.

We determined the sensitivity and specificity of Liverpool and Naranjo tools for predicting ADRs with scores ranging from “Possible” to “Definite” were considered positive and “Unlikely/Doubtful” as negative for ADR.

Results

A total of 924 charts were reviewed and 529 charts contained objective findings to support or against the diagnosis of ADR. The participant age ranged from 1 month old to 93 years. These results were confirmed by laboratory or clinical (re-challenge) testing as noted in 529 cases. Liverpool causality tool had sensitivity (SN) of 97.2% ± 2.4% and specificity (SP) of 2.3% ± 1.57%. The positive (PPV) and negative predictive values (NPV) were 34.1% and 61.5%, respectively. Naranjo scale had SN of 81.2% ± 5.69% and SP of 13.2% ± 3.56%. PPV and NPV were 32.7% and 57.5%, respectively.

Conclusion

The Liverpool causality assessment tool is a more sensitive tool than the Naranjo scale in the assessment of possible ADRs but both tools have poor specificity. The Liverpool tool can be a useful screening tool in a primary care setting where other tests may not be readily available for ADR assessment. However its low PPV and NPV should direct clinicians to pursue further testing when their clinical impression is that the event of interest is an ADR.

References

Naranjo CA, Busto U, Sellers EM, et al. (1981) A method for estimating the probability of adverse drug reactions. Clin Pharmacol Ther 30:239-45

Gallagher RM, Kirkham JJ, Mason JR, et al. (2011) Development and inter-rater reliability of the Liverpool adverse drug reaction causality assessment tool. PLoS One ;6:e28096

Rieder MJ. (2012) New ways to detect adverse drug reactions in pediatrics. Pediatr Clin North Am. 59:1071-92