#5450 EXTERNAL VALIDATION OF A RISK PREDICTION TOOL IN DETERMINING EARLY DEATH AMONG ELDERLY PATIENTS INITIATED ON DIALYSIS

Abstract

Abstract Background and Aims The increase in the senior population is one of the factors that has been attributed to the rise in the number of dialysis patients in the Philippines. The elderly generally has an increased risk for poor dialysis-associated outcomes, and studies have been conflicting if they have better survival with initiation of dialysis as compared to conservative management. The decision-making process in this regard for clinicians, patients, and their families may be ameliorated by the Thamer Risk Scoring tool which uses simple variables such as age, albumin level, nursing home residence, need for assistance in daily living, cancer, heart failure, and hospitalization to predict early mortality in elderly patients after initiating dialysis [1]. This tool has not been validated in the Philippine setting, hence this study aimed to validate the prognostic property of the Thamer Risk Scoring tool in predicting early death among elderly patients initiated on dialysis in a tertiary institution, in Davao City, Philippines. Method A retrospective cohort research design was used. All the charts of patients diagnosed with end-stage renal disease aged 60 years and above, who were admitted and initiated on either hemodialysis or peritoneal dialysis within a 4-year period from 2018 to 2021 were reviewed. Patients with periods of dialysis lasting less than 90 days followed by recovery of kidney function, and patients with lacking data to meet all pre-requisite variables for the risk scoring were excluded. Results A review of 1,583 patient charts was done, with 1,402 omitted based on the exclusion criteria. A total of 181 patients were included in the study. Most of the patients were male (56.35%), and initiated on hemodialysis (97.24%). The mean age of the population was 67.34 years with majority in the 60–69 year age group (66.3%). Those who survived had a mean age of 66.53 years and those who expired had a mean age of 68.63 years. All-cause mortality was 38.12% at 3 months, and 3.57% at 6 months after dialysis initiation. The all-cause mortality at 3 months in this study was three times higher than the Thamer study with a mortality of 12.5%. Patients with heart failure (p-value = 0.007), cancer (p-value = 0.040), asthma (p-value = 0.040), and need for assistance in daily living (p-value = 0.022), had significantly higher mortality rates, while patients who had higher albumin (p-value = 0.010), and creatinine (p-value = 0.014) had higher survival rates. Similar to the Thamer study, this study also showed that the tool becomes less sensitive and more specific as the scores increased for the 3-month, 6-month, and over-all mortality groups. The accuracy of the tool rises with a higher score as well. In the Thamer study, their validation cohort had an area under the receiver operating characteristic curve (AUROC) = 0.691 for their risk scoring tool while this study showed an AUROC = 0.6245, 0.2847, and 0.5993 in the 3-month, 6-month, and over-all mortality groups respectively which does not show statistical strength. Conclusion Patients with older age, comorbidities such as heart failure, cancer, asthma, and those who need assistance in daily living had higher mortality rates. Despite the poor performance of the Thamer risk scoring tool in predicting mortality in terms of AUROC in this study, the sensitivity, and specificity had a similar performance in the Thamer study. Furthermore, the consistent trend of accuracy of predicting mortality in these groups reflect that a higher score portends to a higher accuracy in predicting mortality.