Mortality analysis in hip fracture patients: implications for design of future outcome trials

N. B. Foss1,* and H. Kehlet2

1 Departments of Anaesthesia and Orthopaedic Surgery and 2 Department of Surgical Gastroenterology, Hvidovre University Hospital, Copenhagen DK-2650, Denmark

* Corresponding author. E-mail: nicolai.bang.foss{at}hh.hosp.dk

Accepted for publication September 8, 2004.


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Introduction. Patients with hip fractures are usually frail and elderly with a 30-day mortality in excess of 10% in European series. Perioperative morbidity is often multifactorial in nature, and unimodal interventions will not necessarily decrease mortality. The purpose of this prospective study was to analyse causes of mortality, and thereby the potential and limitations to decrease mortality after hip fracture surgery.

Methods. 300 consecutive, unselected hip fracture patients were treated in a multimodal rehabilitation programme with continuous perioperative epidural analgesia and anaesthesia, early surgery, standardized fluid and transfusion therapy, enforced oral nutrition and early mobilization and physiotherapy. All deaths within 30 days of surgery or during primary hospitalization were analysed and classified according to whether death was unavoidable, probably unavoidable, or potentially avoidable.

Results. Thirty-day mortality was 13.3% (40 patients) and the total perioperative mortality was 15.6% (47 patients). Death was definitely unavoidable in 28%, probably unavoidable in 15%, and in theory potentially avoidable in 57%. In the patients where death was potentially avoidable, active care was curtailed in 16 of 27 (59%) patients.

Conclusion. About a quarter of the total mortality in hip fracture patients is definitely unavoidable, and death is probably only avoidable in about half of the unselected patients. These results have important implications for the design of future outcome studies, which should focus on other relevant outcomes than mortality per se.

Keywords: anaesthesia, audit ; anaesthetic techniques, epidural ; analgesia, techniques, extradural ; complications ; surgery, orthopaedic


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Patients with hip fractures are usually frail and elderly, with more than 30% being 85 yr or older.1 2 In European series, hip fracture patients have a high perioperative mortality and morbidity with a 30-day mortality in excess of 10%,1 and a 1-yr mortality of more than 25%.3 Despite the advances in surgical and anaesthetic techniques during the last 20 yr a decrease in mortality after surgical repair of hip fractures has not been observed.1 Regional anaesthesia with neuraxial blockade has been shown to decrease both postoperative mortality and morbidity in lower body procedures,4 but a meta-analysis of the available studies of regional vs general anaesthesia in hip fracture patients has only shown a non-significant, trend towards decreased mortality.5 Perioperative morbidity in surgical procedures is often multifactorial in nature, and unimodal interventions such as regional anaesthesia or analgesia will not necessarily decrease mortality,6 which may apply in particular to patients with hip fracture.

In order to reduce surgical morbidity the concept of multimodal rehabilitation has been proposed.7 The concept has been evaluated in several surgical procedures as ‘fast track rehabilitation programmes’. The purpose of the current study has been to evaluate the potential possibilities of reducing the postoperative mortality of hip fracture patients within the setting of a well-defined rehabilitation programme including perioperative continuous epidural analgesia, taking preoperative conditions and life expectancy into consideration.


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
From September 2002 until July 2003, 300 consecutive unselected patients from the urban part of Copenhagen were admitted with primary hip fracture to our special hip fracture unit and their outcome was audited prospectively. All patients received a well-defined multimodal fast-track rehabilitation regimen. Patients were admitted to the unit regardless of whether they sustained the fracture at home or during hospitalization for other illnesses. The audit protocol was approved by the local ethics committee.

Within 7 h (median) after arrival, a preoperative continuous lumbar epidural analgesia was instituted. Patients were treated immediately upon arrival at the emergency room according to a standardized fluid and transfusion protocol (re-hydration fluid 20 ml kg–1 and transfusion if haemoglobin was less than 6.0 mmol litre–1). Delay of surgery because of other medical optimization was generally avoided. Anaesthesia was primarily epidural (see Table 1) combined or replaced with general anaesthesia only if required for concomitant upper limb surgery, insufficient epidural blockade or in the case of the epidural not being technically possible or contraindicated as a result of pre-admission anticoagulative therapy.


View this table:
[in this window]
[in a new window]
 
Table 1 Patient characteristics. Values are presented as number of patients (percentage) for nominal data and as median (interquartile range) for data in scales

 
Postoperative epidural analgesia, with local anaesthetics and low-dose morphine (bupivacaine 1.25 mg ml–1 and morphine 0.05 mg ml–1, infusion rate 4 ml h–1), was continued until the fourth postoperative day. In addition, the patients received balanced analgesia with oral paracetamol 1 g x4 and rofecoxib 25 mg x1. The patients were mobilized on the day of the operation with an intensive physiotherapy programme, comprising two daily sessions. All patients received supplemental oxygen when supine in the entire perioperative period, as well as prophylactic antibiotics and low-dose low-molecular weight heparin (enoxaparin 40 mg–1). Preoperative fasting was kept at 6 h for solids and 2 h for fluids, with enforced perioperative oral nutrition and hydration, including supplementary protein drinks.

Patients were in principle rehabilitated in and discharged directly from the orthopaedic ward, except for patients admitted from nursing homes or other hospital wards who were discharged to their previous institution for further rehabilitation when they were deemed medically stable, usually on the third postoperative day.

A database was instituted on all patients admitted to the unit comprising information on preoperative functional level expressed by the New mobility score reflecting functional status (0 meaning total immobility and 9 indicating ability to go shopping unaided),8 mental status, medical conditions, medications, ASA classification, anaesthesia and surgery as well as postoperative complications and discharge status.

Perioperative death was defined as death occurring during primary hospitalization, independent of duration of stay, or within 30 days post-surgery irrespective of hospitalization status. All perioperative deaths were analysed based on information from the Danish personal register on all patients. In the case of death after discharge, the death certificate was scrutinized. All causes of deaths were classified as being either a result of pre-fracture terminal disease, pre-fracture acute medical illness, postoperative complication. In addition, it was noted if the patients had a notation in their charts on restrictions in the level of active therapy. The charts were reviewed independently by the two authors, and deaths were categorized according to consensus. Mortality was classified as definitely unavoidable regardless of the perioperative care regimen, if it was a result of either pre-fracture intractable disease with predictable short life expectancy (<1 month), regardless of the specific diagnosis or the patient refusing relevant postoperative care such as nutrition, fluids, and rehabilitation. A death was termed probably unavoidable regardless of perioperative care regimen if it was clearly a result of a severe pre-fracture acute illness. The remaining deaths—those who were related to postoperative complications—were classified as being potentially avoidable (in theory) by the application of a known/unknown treatment. These patients were further subdivided into patients who did or did not receive maximum perioperative care postoperatively. Patients receiving maximum perioperative care were those who did not at any point have their active therapy restricted by the attending medical staff.


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Perioperative mortality is summarized in Figure 1. Thirty-day mortality was 13.3% (40 patients) and the total perioperative mortality was 15.6% (47 patients). Patient characteristics and anaesthesia data in survivors and non-survivors are presented in Table 1.



View larger version (14K):
[in this window]
[in a new window]
 
Fig 1 Perioperative mortality.

 
The potential relationship between death and perioperative regimen is summarized in Figure 2. Forty-three per cent (CI 95% 0.28–0.57) of all deaths were either absolutely or probably unavoidable regardless of the perioperative regimen. Of those whose deaths were definitely unavoidable, 11 patients had incurable cancer and two refused relevant postoperative treatment. Pathological fracture was only present in two of these patients. Patients with pre-fracture acute illness had either pneumonia or acute incompensated heart failure. Fifty-seven per cent (CI 95% 0.43–0.72) of all deaths were in theory potentially avoidable by an improved (ideal) perioperative regimen. However, a restriction upon the level of active care before the occurrence of death was noted in the charts in 16 of these 27 patients. The reasons for curtailing active care were primarily based on ethical considerations comparing pre-fracture level of function and previous and current health status with previous and sometimes present level of mental function. The decision to curtail active care was taken by the attending medical staff, and not influenced by the authors. Five of the 16 patients who had active care curtailed were triaged for the intensive care unit but not admitted on account of their physical and mental status; two patients who received maximum care were admitted. Thus, a total of 11 of the 47 patients (23%, CI 95% 0.11–0.36) who died postoperatively received maximum perioperative care, and therefore theoretically may have survived with any future ‘ideal’ perioperative treatment regimen.



View larger version (20K):
[in this window]
[in a new window]
 
Fig 2 Mortality related to cause.

 
Death was primarily caused by pneumonia, cancer, or cardiovascular disease. The causes of death are summarized in Table 2. Patients died median 11 (interquartile range 6–18) days after the operation, two patients died more than 100 days postoperatively while still hospitalized after initial admission.


View this table:
[in this window]
[in a new window]
 
Table 2 Classification of patients according to potential for avoiding mortality. Values are presented as number of patients (percentage) for nominal data and as median (interquartile range) data in scales

 

    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
The present study shows that 25–50% of the total mortality in hip fracture patients is probably unavoidable. However, of the remaining patients where death is potentially avoidable, more than half received less than maximum perioperative care, as a result of factors which were not directly part of or necessarily modifiable by a perioperative regimen. Therefore, only about 25% of the perioperative mortality was potentially susceptible to any future interventions. These results have important implications for the design of interventional studies, with the explicit goal of reducing mortality in populations of unselected hip fracture patients. Thus, perioperative regimens comprising regional anaesthetic and analgesic techniques could potentially be feasible methods of reducing postoperative mortality, but if the majority of postoperative deaths is largely dependant on the pre-fracture morbidity and restrictions in level of active care, then the potential effects of regional anaesthesia and analgesia will hardly be reflected in mortality unless a subgroup analysis is performed in large numbers of patients. This has not been the case in previous randomized trials6 or meta-analysis.5

The present study has limitations, as it is a prospective single centre descriptive cohort. On the other hand the mortality rate is comparable with that reported from other series in European countries.1 Reports of very low postoperative mortality rates are often associated with selective patient intake, exclusion of patients with pathological fractures, exclusion of patients sustaining fracture while in hospital, exclusion of patients receiving conservative fracture treatment, and finally insufficient follow-up—especially after early discharge or transfer to secondary rehabilitation units. The long-term mortality after admission to hospital with hip fracture has been shown previously to be causally related to the fracture in 17–32% of all deaths in one study10 and 24–66% in another.11 A study of 30-day mortality only found the diagnosis of fractured hip anywhere on the death certificate in 43% of patients,12 also suggesting that pre-fracture morbidity is considered to be the major determinant of perioperative mortality. The methods used in our study to discriminate avoidable from unavoidable deaths are obviously open to discussion, since it is not based on a specific objective measurement, but instead on an evaluation of each case. The classification was based on whether the patient was terminal or had acute severe illness, before they suffered their fracture, and not upon any specific diagnosis. Also, no evaluation was made of whether the specific cause of death could have been avoided by a further improved care plan. Thus, the group of patients whose death was defined as potentially avoidable, should only be considered avoidable in theory by the future application of some perioperative ‘magic bullet’ intervention, and only so because their death was not assumed to be determined by pre-fracture morbidity. The fatalities classified as definitely unavoidable were reflected in a very low mobility score8 and a high rate of dementia (Table 2). Whether the deaths of two patients that refused active care—the charts explicitly stated that the patient had a wish to die—could/should be avoided is primarily ethical. The seven patients classified as ‘probably unavoidable’ all had serious acute medical conditions leading to their fracture, not the other way round. As such these patients theoretically could be saved by optimized medical care, but the chances for success were considered very low, and again based on ethics and indications for transfer to intensive care unit.

The most critical part of this study is the curtailment of maximum perioperative care by the attending medical staff in so many of the patients. This issue is relevant in hip fracture patients who suffer from many preoperative mental and physical problems; the issue of reasonable level of active care in any given patient will always be open to debate. Thus, in most cases, the decision to curtail active care was taken postoperatively by the attending orthopaedic surgeon, on ethical grounds related to the current or pre-fracture mental or physical condition of the patient. However, it is surprising, and food for thought, that the group receiving maximum care apparently was not different from those who had active care curtailed, regarding functional level, cognitive dysfunction, and place of residence (Table 2). The ethics of operating on a patient, and then not giving maximum perioperative care is debatable, but it must be taken into consideration that patients treated conservatively have an extremely high mortality.9 The possibility of reducing the mortality in the category of patients who received curtailment in active care is highly dependant on attitudes on aggressive postoperative care in very old patients with serious postoperative complications, rather than the perioperative regimen per se.

Outcome studies in hip fracture patients have until now mainly addressed the issues of type of anaesthesia, fluid therapy, and nutrition.52224 Postoperative epidural analgesia has been shown to reduce both pulmonary and cardiac complications in various surgical procedures,1314 but even large-scale studies have failed to show a significant reduction in mortality in specific procedures, which often is attributed to insufficient study power. The available evidence for perioperative continuous epidural analgesia in hip fracture patients in the English language literature is scarce, with only two controlled randomized studies directly addressing the issue. One recent study has shown decreased preoperative cardiac morbidity with a preoperative continuous epidural analgesia.15 Another study of postoperative epidural analgesia showed significant reduction in cumulated perioperative cardiac ischaemia, but not mortality.16 Since hip fracture patients are frail and have several co-morbidities, the advantageous physiological effect of regional anaesthesia and analgesia may only have clinical implications in a restricted number of patients, and the potential advantages of the technique, may be obscured by other perioperative issues. In a recent meta-analysis of general vs regional anaesthesia5 the three largest trials included in the analysis only exclude patients with pathological fracture, contraindications to neuroaxial anaesthesia or age less than 50 yr in one study,17 pathological fracture, multiple trauma, arthroplasty, and age less than 55 yr in another18 and no exclusion criteria in the third study.19 No studies concerning other aspects of perioperative pathophysiology such as thromboembolic complications, stress response, nutritional intake, mobilization and rehabilitation, sleep disorders, or quality of life with continuous regional analgesia techniques compared with systemic opioids are available in this group of patients.

Recent studies have addressed other aspects of perioperative care in the hip fracture population. Two well-conducted studies investigated intra-operative volume optimization as a unimodal intervention, but did not find any significant difference in morbidity or mortality, even when the data were combined as a meta-analysis.2022 In another study a preoperative fluid bolus of colloid vs saline, had no effect on mortality.23 Optimized perioperative nutritional intake only led to a non-significant reduction of morbidity and mortality, but the studies give limited information on other perioperative interventions thereby precluding final interpretation.24 Antibiotic prophylaxis was in a recent meta-analysis shown to reduce postoperative wound and urinary tract infections, but without a reduction in mortality.25

In conclusion, future outcome trials addressing mortality in hip fracture patients must take the present study into account in the design of the trial. If the trial uses a non-selected group of patients, implementation of any therapy will most probably not have any effect on mortality, as only 25–60% of the mortality will be potentially susceptible to the intervention. Future studies on the potential benefits of anaesthetic and analgesic techniques in hip fracture patients should therefore focus on outcome that may be susceptible to the specific intervention, which could include rehabilitation outcomes, as has been demonstrated with regional analgesia techniques in elective orthopaedic procedures.2628


    Acknowledgments
 
This work received financial support from Apotekerfonden af 1991, IMK fonden and The Danish Research Council (no. 22-01-0160).


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
1 Roberts SE, Goldacre MJ. Time trends and demography of mortality after fractured neck of femur in an English population, 1969–98: database study. Br Med J 2003; 327: 771–5[Abstract/Free Full Text]

2 Sharrock NE. Fractured femur in the elderly: intensive perioperative care is warranted. Br J Anaesth 2000; 84: 139–40[Free Full Text]

3 Heikkinen T, Parker M, Jalovaara P. Hip fractures in Finland and Great Britain—a comparison of patient characteristics and outcomes. Int Orthop 2001; 25: 349–54[CrossRef][ISI][Medline]

4 Rodgers A, Walker N, Schug S et al. Reduction of postoperative mortality and morbidity with epidural and spinal anaesthesia: results from overview of randomised trials. Br Med J 2000; 321: 1–12[Abstract/Free Full Text]

5 Urwin SC, Parker MJ, Griffiths R. General versus regional anaesthesia for hip fracture surgery: a meta-analysis of randomized trials. Br J Anaesth 2000; 84: 450–5[Abstract]

6 Holte K, Kehlet H. Postoperative epidural analgesia and outcome—a research agenda. In: Tramer MR, eds. Evidence-Based Resource in Anaesthesia and Analgesia, 2nd edn. London: BMJ Books, 2003; 155–79

7 Kehlet H, Wilmore DW. Multimodal strategies to improve surgical outcome. Am J Surg 2002; 183: 630–41[CrossRef][ISI][Medline]

8 Parker MJ, Palmer CR. A new mobility score for predicting mortality after hip fracture. Br J Bone Joint Surg 1993; 75: 797–8[ISI][Medline]

9 Jain R, Basinski A, Kreder HJ. Nonoperative treatment of hip fractures. Int Orthop 200; 27: 11–17

10 Kanis J, Oden A, Johnell O, De Laet C, Jonsson B, Oglesby AK. The components of excess mortality after hip fracture. Bone 2003; 32: 468–73[CrossRef][ISI][Medline]

11 Parker MJ, Anand JK. What is the true mortality after hip fractures? Public Health 1991; 105: 443–6[ISI][Medline]

12 Goldacre MJ, Roberts SE, Yeates D. Mortality after admission to hospital with fractured neck of femur: database study. Br Med J 2002; 325: 868–9[Free Full Text]

13 Beattie WS, Badner NH, Choi P. Epidural analgesia reduces postoperative myocardial infarction: a meta-analysis. Anesth Analg 2001; 93: 853–8[Abstract/Free Full Text]

14 Ballantyne JC, Carr DB, deFerranti S, et al. The comparative effects of postoperative analgesic therapies on pulmonary outcome: cumulative meta-analyses of randomized, controlled trials. Anesth Analg 1998; 86: 598–612[Abstract]

15 Matot I, Oppenheim-Eden A, Ratrot R, et al. Preoperative cardiac events in elderly patients with hip fracture randomised to epidural or conventional analgesia. Anesthesiology 2003; 98: 156–63[CrossRef][ISI][Medline]

16 Scheinin K, Virtanen T, Kentala E, et al. Epidural infusion of bupivacaine and fentanyl reduces perioperative myocardial ischaemia in elderly patients with hip fracture–a randomized controlled trial. Acta Anaesthesiol Scand 2000; 44: 1061–70[CrossRef][ISI][Medline]

17 Valentin N, Lomholt B, Jensen JS, Hejgaard N, Kreiner S. Spinal or general anaesthesia for surgery of the fractured hip? Br J Anaesth 1986; 58: 284–91[Abstract]

18 Davis FM, Woolner DF, Frampton C, et al. Prospective, multi-centre trial of mortality following general or spinal anaesthesia for hip fracture surgery in the elderly. Br J Anaesth 1987; 59: 1080–88[Abstract]

19 McKenzie PJ, Wishart HY, Smith G. Long-term outcome after repairs of fractured neck of femur. Comparison of subarachnoid and general anaesthesia. Br J Anaesth 1984; 56: 581[Abstract]

20 Sinclair S, James S, Singer M. Intraoperative intravascular volume optimisation and length of hospital stay after repair of proximal femoral fracture: randomised clinical trial. Br Med J 1997; 315: 909–12[Abstract/Free Full Text]

21 Venn R, Steele A, Richardson P, Polniecki J, Grounds M, Newman P. Randomised controlled trial to investigate the influence of fluid challenge on duration of hospital stay and perioperative morbidity in patients with hip fractures. Br J Anaesth 2002; 88: 65–71[Abstract/Free Full Text]

22 Price J, Sear J, Venn R. Perioperative fluid volume optimization following proximal femoral fracture. Cochrane Database Syst Rev 2002; (1)

23 Parker MJ, Griffiths R, Boyle A. Preoperative saline vs gelatin for hip fracture patients; a randomized trial of 396 patients. Br J Anaesth 2004; 92: 67–70[Abstract/Free Full Text]

24 Avenell A, Handoll HHG. Nutritional supplementation for hip fracture aftercare in the elderly (Cochrane review). The Cochrane Library 2004; (1)

25 Southwell-Keely JP, Russo RR, March L, Cumming R, Cameron I, Brnabic AJM. Antibiotic prophylaxis in hip fracture surgery: a metaanalysis. Clin Orthop 2004; 419: 179–84[Medline]

26 Capdevila X, Barthelet Y, Biboulet P, et al. Effect of perioperative analgesic technique on the surgical outcome and duration of rehabilitation after major knee surgery. Anesthesiology 1999; 91: 8–15[ISI][Medline]

27 Chelly JE, Greger J, Gebhard R, et al. Continuous femoral blocks improve recovery and outcome of patients undergoing total knee arthroplasty. J Arthroplasty 2001; 16: 436–45[CrossRef][ISI][Medline]

28 Singelyn FJ, Deyaert M, Joris D, Pendeville E, Gouverneur JM. Effects of intravenous patient-controlled analgesia with morphine, continuous epidural analgesia, and continuous three-in-one block on postoperative pain and knee rehabilitation after unilateral total knee arthroplasty. Anesth Analg 1998; 87: 88–92[Abstract]