J.A. Patel1, J.A. Flippin1, E.J. Medenblik1, F.A. Bokhari1 1Children Hospital Of Illinois / OSF Saint Francis Medical Center, Surgery, Peoria, IL, USA
Introduction: The Model for End Stage Liver Disease (MELD) Score has become the gold standard for evaluating the severity of primary cirrhotic liver disease. Until recently, it had only been used as a predictor of all-cause mortality and not perioperative mortality. Realizing the metabolic derangements accounted for by MELD occur in other disease processes, recent work has validated its use for predicting mortality beyond those with primary cirrhotic liver disease. Since these metabolic derangements may be modifiable, risk reduction analysis can be performed for optimization of reduced MELD scores. Unreported in the current literature is the use of MELD to predict the odds of mortality in patients undergoing intra-cranial or spinal procedures and to evaluate the mortality benefits of optimizing MELD scores.
Methods: The NSQIP data from years 2017-2022 (exclusive of 2020 due to non-standard surgical practices dominant during the SARS-CoV-2 pandemic) was queried for all patients undergoing intracranial procedures and those undergoing spinal fusion or laminectomy. Extensive demographic information was collected including patient characteristics and comorbidities. MELD was calculated using the MELD 3.0 formula. The outcome variable measured was mortality. Chi-square regression was used to calculate the odds of mortality at each MELD score compared to all other MELD scores. Each MELD score was successively and independently used as a reference category to compile comparative odds between all possible MELD score combinations.
Results:
The dataset included 108,886 patients for whom all necessary variables were valid. In the dataset, patients who died were significantly more likely to be older, male, of non-white race, have a higher total MELD score, have a lower preoperative albumin and sodium, and higher INR and creatinine, indicating that the components of MELD were all individually significant. Patients were also more likely to have a higher ASA score, and have hypertension, diabetes, CHF, COPD, ESRD, and a functional status of partially or fully dependent.
Compared to a normal MELD of 6, the odds ratio of mortality for each higher MELD number was significantly higher ranging from an odds ratio of 1.289 for MELD of 7 to an odds ratio of 23.359 at MELD of 31. Above MELD of 31, the number of patients was insufficient to calculate odds accurately.
When evaluating the potential risk reduction of optimizing MELD score, significant benefits were found within the range of MELD numbers for which adequate power existed (up to MELD 32). This showed, for instance, that optimizing a patient from MELD 12 to MELD 11 conferred a 25.7% reduction in the odds of death and reducing MELD from 20 to 17 reduced risk by 53.2%.
Conclusion: MELD is a useful metric for predicting mortality in patients undergoing intra-cranial and spinal surgeries. It can also be used to estimate the mortality benefit conferred by optimizing a patient’s MELD score.