Reporting functional status in UNOS: The weakness of the Karnofsky Performance Status Scale

There is an emerging body of evidence that functional status is a critical determinant of transplant-related outcomes in patients with cirrhosis. Various measures of functional status have been used in the transplant hepatology literature.^1–5 Most recently, Dolgin et al.⁶ reported that the Karnofsky Performance Status (KPS) scale was associated with a nearly 2.5-fold increased odds of 30-day mortality after liver transplantation in a US national cohort of 24 505 liver transplant recipients from 2006 to 2011. While this is one of the first studies to investigate the impact of pretransplant physical function on post-transplant outcomes, we feel compelled to point out some key limitations of the KPS.

One of the first scales of functional status, the KPS was developed in 1949 by Karnofsky et al.⁷ to gauge patients’ potential for benefit from chemotherapy based on their ability to perform daily activities and the level of assistance required to perform these tasks. Its strength is that it is simple to use as a one-time assessment by either the patient or the provider, but its subjective nature leaves it vulnerable to bias. Although the interobserver agreement has not been evaluated in patients with cirrhosis, prior studies in the oncology literature report variable interobserver agreement^8,9; in particular, agreement among observers is often higher among a population that has a good overall performance status than among a population with a more heterogeneous performance status. As such, the KPS has been criticized for poor sensitivity at the lower end of its scale where a wider spread of performance status is observed.¹⁰ Longitudinal assessments of functional status further amplify these weaknesses of the KPS, where it displays poor prognostic value¹¹ and expected fluctuations.¹²

The limitations of this subjective scale are readily apparent when applied to the liver transplant population. In an analysis of adults listed for liver transplantation from 2008 to 2013, we observed that decreasing physical function, based on the KPS scale, was strongly associated with UNOS region (Wald test P<.01), center size (per 100 patients increase, OR 1.08, P<.01), and market competition (2–3 centers/donor service area [DSA]: OR 1.27, P<.01; 4+ centers/DSA: OR 1.62, P<.01), even after adjustment for patient-level factors that we would expect to be associated with physical function such as MELD score, Child-Pugh score, hepatocellular carcinoma, age, and intensive care unit status.¹³ Our data suggest that assessment of physical function of liver transplant candidates by the KPS is biased by factors that would not be expected to be clinically associated with physical function.

The value of Dolgin et al.’s work to the field is in reinforcing the importance of the concept of physical function to our patients undergoing liver transplantation. However, given the implications of these data to the decision to offer transplant to a patient in need, we urge more objectivity in assessments of functional status in the UNOS registry than the KPS can offer. We should use these data to fuel the quest for such metrics in liver transplantation and to justify their implementation in clinical practice and transplant decision-making.