In-hospital mortality does not increase in patients aged over 85 years after hip fracture surgery. A retrospective observational study in a Japanese tertiary hospital

Hip fracture is a common and serious orthopedic injury among the geriatric population, necessitating surgical treatment. Reports from the UK and US indicate that 30-day postoperative mortality and morbidity are still high, at 6.5 and 5.2%, respectively [1, 2]. Although age is one of the significant predictors for increased mortality, it may actually be a surrogate measure for other variables, especially since frailty or reduced functional capacity rather than chronological age predict postoperative mortality after surgery overall [3–5]. Causes and pattern of death after hip fracture surgery are not well understood, and the need to improve postoperative mortality after hip fracture is clear.

In this study, we used the database of hip fracture patients in our hospital to analyze causes and pattern of death. We hypothesized that if age is a clinically significant factor for mortality after hip fracture surgery, there should be a difference between groups when stratified by chronological age.

Iwaki Kyoritsu General Hospital Ethics Committee (the reference number H28-9) approved the study plan. We queried the electronic hospital records of in-patients aged over 75 years who had undergone hip fracture surgery from the start of 2010 to the end of August 2016 in our hospital, Iwaki Kyoritsu General Hospital. The ethics committee waived to obtain individual informed consent to participate in the study from patients.

This hospital is a tertiary hospital covering a population of 348,000 in an area of 1232 km² on the main island of Japan. The extracted data included patient ID, age, gender, location of fracture, ASA score, types of anesthesia, durations of anesthesia and surgery, days of hospital stay after surgery, and outcomes at hospital discharge including in-hospital death. The primary outcome was all-cause in-hospital mortality.

The extracted data were divided into two groups based on the patient’s age at the time of surgery: the aged group (aged less than 85 years) and the advanced age group (aged 85 years or older), and we compared patient characteristics and management variables and discharge disposition between the two groups. We reviewed the electronic records of patients who died in the hospital following surgery to identify the cause of death, which was made from the clinical diagnosis. Currently, post-mortem examination is not usual in Japan when patients die after surgery during a hospital stay.

Statistical analysis was performed using SPSS Ver. 22. Univariate analyses between the aged and advanced age groups were performed using an unpaired t test for continuous variables, and chi-squared test for categorical data. We also used binomial multivariate logistic regression to test whether mortality (a dichotomous-dependent variable) can be predicted based on independent variables: ASA score, age groups, gender, types of anesthesia, and location of fracture. Descriptive data were presented as mean ± standard deviation (S.D.), and a p < 0.05 was considered significant.

In this study, we found, firstly, that in-hospital mortality did not differ between aged and advanced age groups, which was confirmed by a univariate analysis and binominal multivariate logistic regression analysis, secondly, that in-hospital morality was low in both groups as compared to the reported values [1, 2], and, thirdly, that most patients died from advanced cancer or aspiration pneumonia.

There are several explanations for the absence of a difference in mortality between the two groups. First, the in-hospital mortality in this study was less than that previously reported from western countries. Even in those studies, the odds ratio of age among patients aged over 70 years is relatively small [1, 6]. Age should thus be considered as a risk only at the extremes of age, i.e., over 95 years [7]. Second, similarity of ASA-PS scores between the two groups may be an explanation, although we are not aware of the reason why there was no difference in ASA-PS scores between two groups. ASA-PS score, a well-recognized risk factor for postoperative mortality, which was also confirmed in the current study using the binominal multivariate regression analysis, generally increases in parallel with advancing age because the number and severity of co-morbidities also increase with aging. We therefore speculate that the ASA-PS score is the independent risk factor for the postoperative mortality in the patients with hip fracture and that age is a confounding variable having an effect on ASA-PS score. Lastly, longer hospital stays may have contributed to decrease mortality in both groups and hence resulted in the absence of mortality difference between two groups [8]. Many patients stayed in the hospital for more than 20 days after surgery in both groups. The longer hospital stays after surgery may reflect less efficient medical care in our health care system [9]. We also note that acute orthopedic and rehabilitation units are not clearly divided in most Japanese hospitals and postoperative patients often undergo rehabilitation while staying in orthopedic units, which may have better access for medical control.

Although anesthesia-related death is very rare, the best anesthetic techniques for hip fracture surgery are still a matter of debate [10]. In this study, spinal anesthesia was favored over general anesthesia in the advanced age group. We cannot attribute the low in-hospital mortality of the advanced age group to the choice of spinal anesthesia, however, because a mortality benefit with regional anesthesia (epidural or spinal anesthesia) is not proven [11]. We speculate that the attending anesthesiologists intended to avoid the postoperative delirium associated with general anesthesia in the advanced aged patients.

We found that most patients died from aspiration pneumonia or advanced cancer. The former may be regarded as preventable death. Many hip fracture patients show frailty, limited activities of daily living, and difficulty of swallowing on presentation. Visnjevac et al. [12] showed that the sub-classification based on their functional status is useful to predict postoperative mortality among octogenarian ASA-III patients. Preoperative assessment of these functions would be of help identifying patients at risk and thus prevent such complications after surgery. In their retrospective study, Chatterton et al. [1] also showed that the most common cause of death was respiratory infection. Preoperative identification of patients at highest risk for early death would help in tailoring surgical management and improving postoperative outcome, but further study is needed.

This retrospective cohort study from a Japanese tertiary hospital demonstrated that in-hospital mortality after hip fracture surgery is relatively low and that it does not differ between the patients with age of < 85 and with age of ≥ 85 years, suggesting that age is not a clinically significant risk factor for in-hospital mortality. Most causes of in-hospital death were advanced cancer or aspiration pneumonia. The possibility decreasing in-hospital mortality exists in identifying patients at risk of aspiration and preventing it. Further studies are needed to assess and improve swallowing function in those patients.