Cover Page
The following handle holds various files of this Leiden University dissertation:
http://hdl.handle.net/1887/80414
Author: Moerman, S.
Chapter 9
Hip fracture is increasing, and the consequences can be enormous both for patient as individuals and for the healthcare system of a country. This thesis describes several unfavourable outcomes in hip fracture patients and attempts to predict them. The first part is about arthroplasty and how to use it while minimizing the chance of complications. The second part is about identifying patients at risk for adverse outcomes with regard to mortality, delirium, quality of life and function after the hip fracture.
Part I: (Hemi) arthroplasty
Chapter 2 describes 30,830 hip fracture patients registered with the Dutch
131 Summary
in the uncemented group. We conclude that a cemented hemiarthroplasty in elderly patients with a displaced femoral neck fracture results in fewer complications compared to an uncemented hemiarthroplasty.
Part II: Predictors of mortality, delirium, quality of life and daily
life functioning after a hip fracture
In Chapter 4, we describe an external validation study of a prediction model for mortality in the first 30 days after a hip fracture. This model had been developed previously by other authors as an improvement on the widely used Nottingham Hip Fracture Score (NHFS). We tested the score in our cohort of 422 patients of 70 years and older and found good validity: the area under the ROC curve was 0.70 (95% CI 0.60 – 0.79). Therefore, we concluded that this score can be used to identify patients at risk for early mortality.
Chapter 5 describes a prospective cohort study with 378 hip fracture patients. The
department of psychiatry developed a risk model for delirium (RD) based on the published literature at the time. Patients at high risk of delirium according to the RD score were prescribed prophylactic haloperidol (an antipsychotic drug) beginning in 2007. We found no difference in delirium incidence between patients with a hip fracture admitted after 2007 (prophylactic treatment for high-risk patients) and before 2007 (no prophylactic treatment). We concluded that this delirium prevention protocol did not reduce the incidence of delirium.
Chapter 5 further describes the ability of the RD score to identify patients at high risk of delirium. This is possible while all publications we know of have demonstrated that haloperidol does not reduce the incidence of delirium after a hip fracture, therefore the treatment of high-risk patients in our study group has not influenced the delirium incidence in these patients. The area under the ROC curve of the RD score is 0.72 (CI 0.67 - 0.77) (fair). With a cut-off of five, sensitivity of the RD score is 72%, specificity 64%, the negative predictive value 86% and the positive predictive value 42%. Multivariable logistic regression was performed to test the association between the RD score and delirium, length of stay, alternative living situation and mortality. High-risk patients according the RD score had a significant higher incidence of delirium (OR 4.1, CI 2.4-7.0), were more likely to be living at an alternative situation after 3 months (OR 6.6, CI 3.2-13.4) and were less likely to be discharged from the hospital before ten days (OR 1.6, CI 1.0-2.6). The RD score was not associated with mortality. Incidence of delirium in a regression was higher in patients with an RD score ≥5 (OR 4.1, CI 2.4-7.0), male gender (OR 1.9, CI 1.1-3.4)
Chapter 9
and higher age (OR 1.0, CI 1.0-1.1). Thus, the RD score is a useful tool to identify patients with a higher chance of delirium.
The RD score that was introduced in Chapter 5 was further analysed in Chapter
6. Reliability was tested in 102 patients when a second nurse recorded a RD score
for the same patient; intra-class correlation was substantial at 0.77 (0.68–0.84). The optimal cut-off point for balancing sensitivity and specificity was four points (instead of the five points used in Chapter 5). With that new cut-off point, sensitivity was 80% (71–88%), and specificity was 56% (50–62%). Feasibility was tested by controlling all individual RD score sheets for errors compared with the medical chart. In 38 cases, items had a different score than that which would have been concluded from the medical chart (i.e. diagnosis of dementia, age and functional dependence). The clock-drawing test was skipped for 84 patients (22%). Summation of the individual items into the total RD score was incorrect in six patients. Reliability and validity of the individual items of the RD score were analysed. The item ‘use of heroin, methadone or morphine’ was positive in nine patients and ‘daily consumption of four or more alcoholic beverages’ was positive in five patients. Furthermore, the items had low validity (no correlation with delirium); therefore, we propose to remove these items from the RD score. We also propose adding ‘male gender’ and ‘trochanteric fracture’ to the score, as these were risk factors for delirium in a multivariable logistic regression (trochanteric fracture OR 1.79 (CI 1.07–3.01) and male gender OR 1.90 (CI 1.06–3.43)).
In Chapter 7, health- related quality of life (HRQoL), both the physical and mental components, was measured using the SF-12 in 335 hip fracture patients. Both physical and mental HRQoL declined after a hip fracture, but mental HRQoL recovered after one year to pre-fracture values, while physical HRQoL did not. A logistic regression analysis was performed to identify variables that predict this decline in HRQoL. Age younger than 80 years, ASA classification I or II, higher pre-fracture level of mobility, intracapsular pre-fracture and treatment with osteosynthesis (compared to arthroplasty) were associated with greater decline in physical HRQoL. We could not find any risk factors for greater decline in mental HRQoL.
133 Summary
1 year. A logistic regression analysis identified higher age, living with a partner pre-fracture, living at home pre-fracture, walking independently pre-fracture and longer length of hospital stay as risk factors for a larger loss of (i)ADL. Living with a partner pre-fracture and use of walking aids pre-fracture were associated with greater recovery of (i)ADL between 3 and 12 months after the fracture. Correlation between (i)ADL, living situation and the use of walking sticks was measured to gain an understanding of the consequences of lower (i)ADL. Correlation between (i) ADL and living situation was substantial (0.69), as was correlation between (i)ADL and mobility (0.80).
