Author information
- 1Department of Orthopaedic Surgery, Harvard Medical School Orthopaedic Trauma Initiative, Massachusetts General Hospital, Boston, MA, USA.
- 2Department of Surgery, University Medical Center Utrecht, Utrecht, the Netherlands.
Abstract
Background: Liver cirrhosis is associated with osteoporosis, imbalance leading to falls, and subsequent fragility fractures. Knowing the prognosis of patients with liver cirrhosis of varying severity at the time of hip fracture would help physicians determine the course of treatment in this complex patient popultaion.
Questions/purposes: (1) Is there an association between liver cirrhosis of varying severity and mortality in patients with hip fractures? (2) Is there an association between liver cirrhosis of varying severity and the in-hospital, 30-day, and 90-day postoperative complications of symptomatic thromboembolism and infections including wound complications, pneumonia, and urinary tract infections?
Methods: Between 2015 and 2019, we identified 128 patients with liver cirrhosis who were treated for hip fractures at one of two Level I trauma centers. Patients younger than 18 years, those with incomplete medical records, fractures other than hip fractures or periprosthetic hip fractures, noncirrhotic liver disease, status after liver transplantation, and metastatic cancer other than hepatocellular carcinoma were excluded. Based on these exclusions, 77% (99 of 128) of patients were eligible; loss to follow-up was 0% within 1 year and 4% (4 of 99) at 2 years. The median follow-up duration was 750 days (interquartile range 232 to 1000). Ninety-four patients were stratified based on Model for End-stage Liver Disease (MELD) score subgroup (MELD scores of 6-9 [MELD6-9], 10-19 [MELD10-19], and 20-40 [MELD20-40]), and 99 were stratified based on compensation or decompensation status, both measures for liver cirrhosis severity. MELD scores combine laboratory parameters related to liver disease and are used to predict cirrhosis-related mortality based on metabolic abnormalities. Decompensation, however, is the clinical finding of acute deterioration in liver function characterized by ascites, hepatic encephalopathy, and variceal hemorrhage, associated with increased mortality. MELD analyses excluded 5% (5 of 99) of patients due to missing laboratory values. Median age at the time of hip fracture was 69 years (IQR 62 to 78), and 55% (54 of 99) of patients were female. The primary outcome of mortality was determined at 90 days, 1 year, and 2 years after surgery. Secondary outcomes were symptomatic thromboembolism and infections, defined as any documented surgical wound complications, pneumonia, or urinary tract infections requiring treatment. These were determined by chart review at three timepoints: in-hospital and within 30 days or 90 days after discharge. The primary outcome was assessed using a Cox proportional hazard analysis for the MELD score and compensation or decompensation classifications; secondary outcomes were analyzed using the Fisher exact test.
Results: Patients in the MELD20-40 group had higher 90-day (hazard ratio 3.95 [95% CI 1.39 to 12.46]; p = 0.01), 1-year (HR 4.12 [95% CI 1.52 to 11.21]; p < 0.001), and 2-year (HR 3.65 [95% CI 1.68 to 7.93]; p < 0.001) mortality than those in the MELD6-9 group. Patients with decompensation had higher in-hospital (9% versus 0%; p = 0.04), 90-day (HR 3.35 [95% CI 1.10 to 10.25]; p = 0.03), 1-year (HR 4.39 [95% CI 2.02 to 9.54]; p < 0.001), and 2-year (HR 3.80 [95% CI 2.02 to 7.15]; p < 0.001) mortality than did patients with compensated disease. All in-hospital deaths were related to liver failure and within 30 days of surgery. The 1-year mortality was 55% for MELD20-40 and 53% for patients with decompensated disease, compared with 16% for patients with MELD6-9 and 15% for patients with compensated disease. In both the MELD and (de)compensation analyses, in-hospital and postdischarge 30-day symptomatic thromboembolic and infectious complications were not different among the groups (all p > 0.05). Ninety-day symptomatic thromboembolism was higher in the MELD20-40 group compared with the other two MELD classifications (13% for MELD20-40 and 0% for both MELD6-9 and MELD10-19; p = 0.02).
Conclusion: The mortality of patients with preexisting liver cirrhosis who sustain a hip fracture is high, and it is associated with the degree of cirrhosis and decline in liver function, especially in those with signs of decompensation, defined as ascites, hepatic encephalopathy, and variceal hemmorrhage. Patients with mild-to-moderate cirrhosis (MELD score < 20) and those with compensated disease may undergo routine fracture treatment based on their prognosis. Those with severe (MELD score > 20) or decompensated liver cirrhosis should receive multidisciplinary, individualized treatment, with consideration given to palliative and nonsurgical treatment given their high risk of death within 1 year after surgery.
Level of evidence: Level III, therapeutic study.