Reply

TO THE EDITOR:

We appreciate Dr. Eyal Shemesh’s careful review of our article titled “The Stanford Integrated Psychosocial Assessment for Transplant Is Associated With Outcomes Before and After Liver Transplantation.”⁽¹⁾ The data for the tacrolimus coefficient of variation (CoV) is emerging in liver transplantation (LT), and intrapatient tacrolimus variability can be calculated several ways. Several studies have established relationships between the Medication Level Variability Index (MLVI) score, which is calculated as the standard deviation of at least 3 consecutive tacrolimus levels, and important post-LT outcomes⁽²⁾ and found that higher prerejection MLVI scores were associated with rejection in pediatric patients. Supelana et al.⁽³⁾ demonstrated similar findings in adult patients.

In response to the author’s comment, we calculated the MLVI in our study sample and found a median MLVI score of 2.4 (interquartile range, 1.9–3.3). Patients with an elevated Stanford Integrated Psychosocial Assessment for Transplant (SIPAT) score (≥21, indicating higher psychosocial risk) exhibited a higher median MLVI score (2.8) compared with patients with a lower risk SIPAT (2.3; P = 0.002). Moreover, patients with a SIPAT score ≥21 were more likely to have a MLVI score ≥2.5 (59% compared with 42%; P = 0.004) and ≥3.0 (43% compared with 27%; P = 0.003; Fig. 1A). These findings are similar to the results in our article, which demonstrated that patients with a SIPAT score ≥21 had a significantly higher median tacrolimus CoV (0.46 versus 0.36; P < 0.001).⁽¹⁾

FIG. 1.

(A) Relationship between SIPAT and MLVI scores. Higher scores indicate higher psychosocial risk. MLVI score = a standard deviation of at least 3 consecutive tacrolimus levels. (B) Adjusted impact of MLVI score on risk of time to biopsy-proven allograft rejection. Model is adjusted for age, sex, race, liver disease diagnosis, education, and community health score.

Employing a Cox regression analysis adjusting for sociodemographic variables, we found that patients with a MLVI score ≥2.5 had a higher risk of biopsy-proven rejection (hazard ratio [HR], 2.29; 95% confidence interval [CI], 1.14–4.58; Fig. 1B). A cutoff of ≥3.0 produced similarly significant findings (HR 2.14, 95% CI 1.12– 4.11), and a cutoff of ≥2.0 was nonsignificant (HR, 1.64; 95% CI, 0.87–4.63). Tacrolimus CoV did not demonstrate a significant relationship with biopsy-proven rejection.

The relationship we observed between tacrolimus CoV and SIPAT score in our analyses was similar to those observed between the MLVI and SIPAT. However, current analyses revealed a significant relationship between MLVI score and biopsy-proven rejection in our study sample, but not with tacrolimus CoV. This finding could reflect the author’s point that although the tacrolimus CoV corrects for mean tacrolimus level, this correction can underrepresent fluctuation values that are low/closer to zero. Ultimately, both measures can provide valuable insight into medication adherence and risk of rejection, and their differences, advantages, and disadvantages should continue to be examined in future studies.