Keywords
hip injuries - mortality rate - epidemiology
Introduction
Hip fracture is a pathology with significant morbidity and mortality, especially in patients older than 60 years of age.[1] The fatality per year has been classically described as of approximately one third, and, in a hospital in Chile, in a studied that included patiens from 2010 to 2012, a fatality rate of 27% was reported at 1 year between 2010 and 2012.[2] However, there is currently a downward trend throughout the world, reaching 23.3% in Europe, 17.9% in Asia, and 24.9% in the United States.[3]
A modern health system should aim at reducing the incidence and fatality in hip fracture.[4] The latter is a classic health indicator, and corresponds to the number of deceased patients one year after the event divided by the number of patients who sustained a fracture during the study period. There is not much evidence of what happens after one year; however, it is known that, after an osteoporotic fracture, life expectancy decreases compared to that of the general population.[5]
Before planning health interventions, an analysis of the trend over the years of the health indicators related to the pathology is required.[6] The purpose of the present study is to evaluate the trend of fatality at one and two years, the waiting time for surgery, and length of hospital stay for hip fracture between 2002 and 2016 in a Chilean health center. Two hypotheses are raised: first, that, in the period studied, there will be a decrease in the annual fatality and in the waiting time for surgery. Second, a direct relationship is expected between fatality at one and two years after a hip fracture in patients older than 60 years of age.
Materials and Methods
A cross-sectional study was designed and approved by the ethics committee of our institution. All patients older than 60 years of age who were coded at the time of admission with a diagnosis of hip fracture between January 1, 2002, and December 31, 2016, were included. Our hospital codes diseases according to the 10th International Classification of Diseases (ICD-10). A search was made for codes S72.0 (femoral head and neck fracture), S72.1 (pertrochanteric fracture), and S72.2 (subtrochanteric femur fracture). The full name, the national identification number, the date of birth, the date of admission, the date of discharge, the code of the surgery, and the date of the surgery were recorded. All patients for whom we did not have all the information were excluded. Then, following the regular channel through Portal Nacional de Transparencia (Chilean National Transparency Portal), the Registro Civil Nacional (Chilean National Civil Registry) was requested to report if the patients were alive as of June 30, 2019, or, otherwise, inform the date of death.
We used the Spearman correlation to determine the behavior of the variables studied (lethality at 1 and 2 years, waiting time for surgery, and hospital stay) during the period studied (2002 to 2016), using the rho parameter. If the value of rho is positive, it is interpreted as a tendency to increase in the period under study, while if the value of rho is negative, the tendency is to decrease. A probability lower than 0.05 is interpreted as a significant correlation. The Spearman correlation was also used to correlate the length of hospital stay and the waiting time for surgery.
To analyze the association between fatality at one and two years, only patients older than 60 years of age were included, and the Pearson correlation was used. Before the analysis, the normal distribution of both variables was verified. If the results of the Skewness, Kurtosis, and Shapiro-Wilks showed a probability higher than 0.15, the null hypothesis (values follow a normal distribution) was assumed. The results of the Skewness, Kurtosis, and Shapiro-Wilks tests were 0.53, 0.77, and 0.36 for the fatality at 1 year, and 0.69, 0.25, and 0.29 for the fatality at 2 years respectively, so a normal distribution was assumed. Next, a linear regression model was estimated, reporting β0, β1, and R2. For the analysis after the estimation of the linear regression, the assumption of heteroscedasticity, the Cook distance, and the normal distribution of the residuals were checked to validate the regression. The Cook-Weisberg test for heteroscedasticity was used, and values of p > 0.15 were considered acceptable. The Cook distance was calculated, accepting a maximum distance of 0.99. The residuals of the linear regression were analyzed using the Skewness, Kurtosis, and Shapiro-Wilks tests, accepting the normal distribution of the residuals if p > 0.15. For internal validity, a Bootstrap regression with 200 repetitions was estimated. The data was processed using the Stata (StataCorp LP, College Station, TX, US) software, version 15.
Results
We investigated 1,064 admissions for hip fracture, 961 (90.32%) of which were of patients older than 60 years of age, and they were included in the study. The median age at the time of the fracture was 83 years (range: 61 to 105 years; interquartile range [IQR]: 76 to 88 years). In total, 763 (79.90%) cases correspond to female patients. There were 389 (40.47%) cases of intracapsular fracture, and 468 (48.70%) cases of extracapsular fracture; in 104 (10.82%) cases, it was not possible to determine whether the fracture was intra- or extracapsular. The median waiting time for surgery was of 3 days (range: 1 to 42 days; IQR: 2 to 5 days), and the median hospital stay was of 8 days (range: 1 to 105 days; IQR: 5 to 12 days).
During the period of study, there was a decreasing trend decrease in the lethality at 1 (rho = -0.23) and 2 years (rho = -0.29), in the proportion of men (rho = -0.25), in the hospital stay (rho = -0.06), and in the waiting time for surgery (rho = -0.11), without reaching statistical significance ([Table 1]). On the other hand, the proportion of extracapsular fractures (rho = 0.04) and age (rho = 0.09) showed an increasing trend, without reaching statistical significance ([Table 1]). The length of the hospital stay had a strong positive Spearman correlation with the waiting time for surgery (rho = 0.76; p ≤ 0.0001), and the median of the ratio between the waiting time for surgery and the hospital stay was of 0.4 (range: 0.1 to 5.7).
Table 1
|
2002-2016 (rho)
|
p
|
|
Age at the time of hip fracture
|
0.09
|
0.76
|
|
Proportion of men
|
−0.25
|
0.37
|
|
Proportion of extracapsular fractures
|
0.04
|
0.88
|
|
Length of hospital stay
|
−0.06
|
0.93
|
|
Waiting time for surgery
|
−0.11
|
0.87
|
|
Fatality at one year
|
−0.29
|
0.29
|
|
Fatality at two years
|
−0.23
|
0.41
|
The fatality at 1 year was of 0.17 (95% confidence interval [95%CI]: 0.14 to 0.19), and, at 2 years, it was of 0.26 (95%CI: 0.23 to 0.29). Both variables are summarized by year in [Table 2] and [Figure 1]. The Pearson correlation between the fatality at 1 and 2 years was of 0.78 (p = 0.001). The estimation of the linear regression determined a constant (β0) of 0.10 (p < 0.000) and a β1coefficient for fatality at 1 year of 0.89 (p < 0.000), while the value of R2 was of 0.60. The Bootstrap estimate with 200 repetitions showed a constant β0 of 0.10 (p < 0.000) and a β1coefficient of 0.90 (p < 0.000).
Table 2
|
N
|
Fatality at 1 year
|
Fatality at 2 years
|
|
2002
|
75
|
0.24 (0.14–0.34)
|
0.34 (0.23–0.45)
|
|
2003
|
61
|
0.16 (0.07–0.26)
|
0.30 (0.18–0.41)
|
|
2004
|
86
|
0.26 (0.16–0.36)
|
0.28 (0.18–0.38)
|
|
2005
|
70
|
0.20 (0.11–0.28)
|
0.30 (0.20–0.40)
|
|
2006
|
62
|
0.09 (0.02–0.15)
|
0.17 (0.08–0.27)
|
|
2007
|
66
|
0.08 (0.01–0.15)
|
0.20 (0.09–0.30)
|
|
2008
|
52
|
0.17 (0.07–0.27)
|
0.21 (0.10–0.33)
|
|
2009
|
107
|
0.14 (0.07–0.21)
|
0.26 (0.18–0.35)
|
|
2010
|
73
|
0.19 (0.09–0.28)
|
0.28 (0.17–0.38)
|
|
2011
|
56
|
0.14 (0.05–0.24)
|
0.20 (0.09–0.30)
|
|
2012
|
48
|
0.15 (0.04–0.25)
|
0.21 (0.09–0.33)
|
|
2013
|
49
|
0.17 (0.06–0.28)
|
0.24 (0.12–0.37)
|
|
2014
|
39
|
0.23 (0.09–0.37)
|
0.31 (0.16–0.46)
|
|
2015
|
49
|
0.14 (0.04–0.24)
|
0.27 (0.14–0.39)
|
|
2016
|
54
|
0.15 (0.05–0.25)
|
0.27 (0.15–0.40)
|
Fig. 1 Distribution of fatality at one year (black) and two years (gray) throughout the study period (2002 to 2016).
Discussion
The most significant finding of the present study is the confirmation of our second hypothesis. After a hip fracture, the fatality rates at one and two years are strongly correlated, so a constant risk of dying after a hip fracture can be inferred in patients aged 60 years or older. The estimation of the linear regression shows that 60% of the reasons for mortality at 2 years are significantly explained by the fatality at 1 year. Given the aforementioned information, the risk factors for fatality at one year – socioeconomic situation, waiting time for surgery, and comorbidities[7]–determine the risk of dying in a period longer than just one year.
Hip fracture is still a major public health concern around the world. National population studies conducted in Northern Ireland,[8] Korea,[9] Japan,[10] and China[11] show an upward trend in incidence. Osteoporosis, sarcopenia, and falls are the main modifiable risk factors. Frailty in elder adults with hip fracture should put prevention at the forefront of the discussion. In this regard, the treatment of osteoporosis, a silent epidemic,[12] and sarcopenia with muscle training programs[13]
[14] could reduce the incidence of hip fracture.
The fatality rate at our center has not varied throughout the period studied, not fulfilling our first hypothesis. One of the main reasons is that the rate is at the lower limit described in recent reports in the international literature.[15] A study published in 2015[2] in a public hospital in the fifth region of Chile showed a fatality rate of 27 %. Despite the low rate, this analysis should encourage us to make greater efforts to reduce the fatality rate. One of the main modifiable findings of the present study is that 50% of our patients undergo surgery 3 days after admission, which is above the standard recommended by the Scottish national guidelines for the treatment of hip fracture, which is 2 days.[16] On the other hand, since the waiting time for surgery is directly correlated with the hospital stay, reducing it also reduces costs for both the patient and the institutions.
Strategies aimed at increasing the availability of operating rooms for these patients and a multidisciplinary team that focuses on stabilizing the comorbidities of the partients are important to reduce the length of the hospital stay and the fatality rate.[17] A joint effort between the geriatrician, internist, orthopedic surgeon, and anesthesiologist will enable the surgery to be performed in an adequate balance between the minimum waiting time and the stabilization of the patient's comorbidities. A recent meta-analysis by Klestil et al.[18] showed that surgery within 48 hours was associated with a lower risk of mortality at 1 year and fewer perioperative complications. Likewise, Johanssen et al.[19] showed that the absolute perioperative risk of death is lower than 1.5% in patients considered class 4 on the American Society of Anesthesiologists (ASA) classification who underwent surgery for hip fracture.
A recent study[20] showed that the incidence of hip fracture in Chile is of 40 cases per 100 thousand inhabitants. On the other hand, another recent study[22] projected a 27% growth in the incidence of fractures by 2030. The estimate of mortality in Chile in 2017 is around 30%, with an estimated number of deaths per year close to the number of patients who die per year from colon cancer.[20] Adding these arguments to the fact that mortality is not only affected until one year after the fracture, but at least 2 years after it, hip fracture constitutes a serious health problem in Chile. Two aspects must be emphasized: the prevention approach and the opportunity to access surgery, which currently only reaches 80% of the population and with a long waiting time.[20]
[21]
One of the most significant changes in Chile in terms of public health strategies in the last 20 years has been the Garantías Explícitas de Salud (GES, Explicit Health Guarantees), which constitute a set of benefits guaranteed by law, which enable the guarantee of access, opportunity, financial protection, and quality in a limited list of diseases. For example, the inclusion in the GES of acute myocardial infarction reduced the annual mortality due to this pathology.[22] Given the worldwide increase in the incidence of hip fracture and the impact of timely access to surgery, not only in the fatality rate per year, but at least foe two years, hip fracture is a serious candidate to be considered in health strategies in Chile.
A limitation of the present study is that the data prior to 2012 were registered manually; as of 2012, an electronic record system was implemented in our health center. Also, another limitation is that the specific cause of mortality is unknown, so a patient could have a hip fracture, but die of cancer, a pathology not related to the event, or prostration. However, the number of patients studied minimizes this bias. Finally, factors influencing the relationship between the surgery time and hospital stay were not analyzed.
Conclusion
The fatality at one and two years, the waiting time for surgery, and the length of hospital stay have remained constant during 2002 and 2016 in our health center. A strong direct correlation has been found between fatality at one and two years, suggesting a consistently high risk of dying after a hip fracture. A strong correlation was also found between the length of the hospital stay and the waiting time for surgery.
