6431 Comparison of Diagnostic Scoring Systems for Myxedema Coma

Abstract Disclosure: K.S. Wei: None. A. Joyo: None. A.N. Lim: None. T.E. Angell: None. Introduction: Myxedema coma (MC) is a rare condition of severe hypothyroidism with high mortality. Diagnostic scoring systems exist, such as criteria by Popoveniuc et al. (2014) and Chiong et al. (2015), which both include categories of: “highly likely MC” (hlMC), “suggestive of MC” (sMC), and “unlikely MC” (uMC). Because few data support their clinical utility, we performed an analysis comparing these diagnostic criteria and associations with mortality and hospital measures. Methods: Patients with MC were included from a retrospective cohort of hospitalized adult patients with TSH greater than or equal to 10 mIU/L and free thyroxine (fT4) below the lower reference limit. Initial clinical diagnosis of MC was generally based on: hypothyroidism, precipitating illness, hypothermia, and altered mentation. Data collection included patient characteristics, mortality, ICU admission, and hospital length of stay (LOS). Patients were post-hoc categorized based on criteria of Popoveniuc and Chiong. Results: Of 386 initial patients, 25 patients had a clinical diagnosis of MC (median age = 60.1 years [IQR: 53.6-70.0]; 10/25 [40%] female). Median TSH was 58 mIU/L (IQR: 33.8-93.3) and median fT4 was 0.48 ng/dL (IQR: 0.16-0.68). All patients received IV levothyroxine and supportive care. Using Popoveniuc, 20/25 (80%) patients were hlMC and 5/25 (20%) were sMC. In contrast, using Chiong, 11/25 (44%) were hlMC, while 14/25 (56%) were sMC. No patient was considered uMC by either system. 10/25 (40%) patients were hlMC in both systems. 10/25 (40%) were hlMC in Popoveniuc but not Chiong, while only 1/25 (4%) was hlMC in Chiong but not Popoveniuc. Finally, 4/25 (16%) were not hlMC in either system. Mortality occurred in 3/25 (12%) patients, 2 of which were hlMC in Popoveniuc but not Chiong. There was a trend toward longer median hospital LOS in patients identified as hlMC vs not hlMC using Popoveniuc (11 days [IQR: 5.3-37.0] vs. 3.0 days [IQR: 2.0-17.5], p=0.07) but no statistically significant differences in hospital LOS were noted. There were no significant differences in ICU admission or LOS. Conclusion: Existing diagnostic scoring criteria of MC identified most patients but differed in their categorizations. Some cases of clinically diagnosed MC were not identified as hlMC by either scoring criteria. Criteria by Popoveniuc et al. appears to better capture patients most at risk for mortality and prolonged hospitalization. Presentation: 6/2/2024