In the case of a hip fracture, operating on a patient with dementia can be a difficult decision to make. Indeed, surgery exposes patients with dementia to a higher mortality rate, more delirium and postoperative complications, and a greater loss of mobility than patients of the same age without cognitive impairments. For patients with dementia in institutional settings, survival is better for those who undergo surgery than for those who do not. But what about the prognosis of surgery vs no surgery for patients with dementia who live at home?
To answer this question, researchers in the United States conducted a cohort study using Medicare data. This retrospective study included patients aged 66 years and older with dementia who were living at home and not placed in institutions and who had a hip fracture between January 2017 and June 2018. Patients with incomplete observations, particularly regarding the location and type of residence (home/institution), were excluded from the analysis. Fractures were categorized as (i) fractures of the head and neck, (ii) pertrochanteric, (iii) subtrochanteric, and (iv) multiple/complex. The type and severity (mild, moderate, or severe) of dementia were identified using the diagnostic code list in the International Statistical Classification of Diseases, Tenth Revision.
Examining Mortality
The primary outcome was mortality at 30, 90, and 180 days. Secondary outcomes included hospital delirium; the need for subsequent hospitalization (within 6 months) or home health services (within 10 days); and intensive care interventions such as dialysis, intubation, resuscitation maneuvers, mechanical ventilation, or the insertion of a feeding tube.
Postoperative medical facility admissions were distinguished according to whether there was a plan to return home. To compare the surgery and nonsurgery groups, an inverse propensity score analysis was conducted within subgroups determined by fracture type, comorbidities (using the Elixhauser score), the person or entity responsible for admission (ie, physician, clinic, hospital, etc.), dual eligibility (Medicare and Medicaid), place of residence (ie, urban or rural), race, and sex. Dementia severity was estimated using a frailty index on the basis of evaluation tests that were systematically collected in the Medicare database.
Results Favored Surgery
Among 56,209 patients with dementia who were admitted for a hip fracture (73.0% women; mean age, 86.4 ± 7.0 years), 33,142 (59.0%) underwent surgery and 23,067 (41.0%) did not. Among surgically treated patients, 73.3% had fractures of the head or neck of the femur, and 40.2% had moderate to severe dementia. The nonsurgically treated fractures were 78.5% pertrochanteric. Comorbidities were evenly distributed between the two groups.
At 180 days, mortality was 31.8% in the surgery group compared with 45.7% in the nonsurgery group, resulting in a significant reduction in the unadjusted relative risk (RR) for death in favor of surgery (RR, 0.67; 95% CI, 0.60-0.76; P < .001). Among patients with mild dementia and a fracture of the head or neck of the femur, mortality at 180 days was 26.5% among surgical patients and 34.9% among nonsurgical patients (RR, 0.67; 95% CI, 0.60-0.76; P < .001). After the investigators adjusted for risk according to propensity score, the benefit of surgery remained significant at 30, 90, and 180 days, regardless of dementia severity. There was no significant difference in mortality for other types of hip fractures between the surgery and nonsurgery groups, however.
The adjusted RR for in-hospital delirium was 1.23 (P = .008), which was significant for the surgery group, but only for those with moderate to severe dementia. There were also fewer permanent placements (P < .001) among the surgically treated patients, and fewer patients with mild dementia required nurse care at home. There was no difference in resuscitation maneuvers between surgery and nonsurgery patients, whether the dementia was mild or not. For patients with a fracture of the head or neck of the femur, there was no difference in the likelihood of rehabilitation admission within 180 days, whether they were operated on or not.