Abstract
Background
Hip fractures are associated with serious morbidity and mortality. Low haemoglobin at presentation has been shown to be associated with increased mortality in hip fracture patients. This comorbid patient group commonly receives packed red cell blood transfusions during their hospital admission, the impact of which is less clear.
Aims and objectives
We aim to assess the rate, appropriateness and impact of blood transfusions on one-year mortality in hip fracture patients. We also aim to assess the impact of patients taking anticoagulant medications at presentation on the rates of blood transfusions in this patient group.
Methods
A retrospective cohort study of 324 consecutive hip fracture patients. Data was collected from the national hip fracture database, electronic patient records and PACS.
Results
75 patients received a blood transfusion. Receiving a blood transfusion increased absolute risk of one-year mortality by 2.466 (p < 0.05). Adjusted for age, sex, comorbidities, residence prior to admission and time from presentation to surgery increased the risk of one-year mortality was 2.790 (p < 0.05).
28% of patients who went on to receive a transfusion had a haemoglobin of less than 100 g/L at presentation. 94.6% of transfused patients had a pre-transfusion haemoglobin of less than 90 g/L. There was no increased risk of requiring a blood transfusion if anticoagulant medication was being taken at presentation.
Conclusion
Receiving a blood transfusion during an admission for hip fracture carried an increased risk of one-year mortality of almost two and a half times. With appropriate preoperative optimisation, taking an anticoagulant medication at presentation did not increase the risk of requiring a transfusion. Most blood transfusions were administered appropriately using thresholds. Just over a quarter of patients who received a transfusion had an admission haemoglobin of less than 100 g/L, showing it as a poor predictor of blood transfusion requirement during admission.
Keywords
Mr Michael Sean Greenhalgh: Conceptualisation, data collection, data analysis, wrote first draft and final manuscript, Mr Benjamin Thomas Vincent Gowers: Literature search, data analysis, contributed to first drafts and final manuscript, Mr Karthikeyan P Iyengar: Literature search, manuscript writing and editing. Mr Riad F Adam: Supervising consultant for work and approved final draft. All authors approved the final manuscript.
1. Introduction
Neck of Femur (NOF) fracture or Proximal Femoral Fracture (PFF), commonly referred to as a hip fracture, is a common cause for presentation to the Emergency Department (ED) in the United Kingdom. They are defined as a break in the upper region of the femur, between the subcapital region (the area just under the femoral head) and 5 cm below the lesser trochanter.
1
The 2019 annual report from the National Hip Fracture Database (NHFD) reported that in England, Wales and Northern Ireland, 66,313 people presented with NOF fracture. In health care costs, the management of hip fractures accounts for over 1% of the total National Health Service (NHS) budget. They are associated with a total cost to health and social services of over £1 billion per year.
2
The incidence of NOF fractures increases with age, with the most common patient group being elderly, postmenopausal women.
2
,3
Hip fractures are associated with serious morbidity and mortality. Factors affecting outcomes are well documented in literature.4
, 5
, 6
Current National Institute for Health and Clinical Excellence (NICE) and British Orthopaedic Association (BOA) guidelines suggest early operative intervention has been shown to improve outcomes.7
,8
The Nottingham Hip Fracture score is made up of seven independent predictors of mortality; age, sex, abbreviated mental test score (AMTS), low haemoglobin (Hb) on admission, residence pre-fracture, comorbidities and active malignancy in the previous 20 years. The score has been validated for the prediction of mortality at both 30 days and 1-year post injury.9
,10
Whilst low haemoglobin has been shown to be associated with increased mortality in hip fracture patients, the impact of receiving a Packed Red Cell (PRC) blood transfusion is less clear.11
,12
The transfusion of packed red cells is indicated when haemoglobin drops below a threshold. This threshold is dependent on national guidelines. In the United Kingdom the National Institute for Clinical Excellence (NICE) advises the consideration of transfusion at or below 70 g/L, or at or below 80 g/L in the presence of cardiac complications. The Joint United Kingdom Blood Transfusion and Tissue Transplantation Services Professional Advisory Committee (JPAC) advises consideration of transfusion below 80 g/L. Internationally, the American Association of Blood Banks (AABB) guidelines suggest a threshold of 80 g/L for consideration of transfusion in surgical patients.
13
, 14
, 15
We assess the implications and frequency of PRC blood transfusions in hip fracture patients and their effects on one-year mortality.
2. Methods
2.1 Study design and setting
We performed a retrospective cohort study, identifying all consecutive hip fractures presenting to Southport and Ormskirk District General Hospital in the United Kingdom over a one-year period.
2.2 Study approval and statement of ethics
The study protocol was approved by the local review board of the research and clinical effectiveness department. The patients undergoing surgery gave their written informed consent; for patients who lacked capacity, consent was obtained in accordance with the local consenting and ethics guidelines and in discussions with their families.
2.3 Patients
Consecutive hip fracture patients who presented between October 01, 2017 and September 30, 2018 were included.
2.4 Data collection
Data for these patients was collected from the National Hip Fracture Database (NHFD), electronic patients notes (EVOLVE, Kainos UK), Picture Archiving and Communication System (PACS), orthopaedic handover documentation, patient discharge summaries and data from the local hospital transfusion laboratory. For the purposes of our study, patients were divided into two cohorts dependent on if they received a PRC blood transfusion at any point during their admission.
2.5 Statistical analysis
Statistical analysis of collected data was undertaken with Statistical Package for the Social Sciences software (SPSS for mac, build 1.0.0.1461, IBM).
3. Results
3.1 Demographics
A total of 324 patients presented with hip fractures in the one-year study period and were included. No patients were excluded. No patient was lost to follow up.
The mean age of these patients was 83.23 (SD 7.69), 251 (77.4%) were female and 73 (22.5%) were male. The mean American Society of Anaesthesiologists (ASA) grade was 2.8 (SD 0.621). Before their injury, 241 patients (74.4%) lived in their own home, 53 (16.4%) in a residential home and 30 (9.3%) in a nursing home.
The patients were split into two cohorts, those who received a transfusion in their admission and those who did not. Two-sample t-tests assuming equal variances were performed, showing there was no statistically significant difference in patient's age (p = 0.008) or ASA grade (p = 0.047) between cohorts.
The full details of patient demographics of each cohort can be found in Table 1.
Table 1Patient demographics, ASA grade and time to surgery.
| Parameters | Did Not Receive Blood Transfusion (n = 249) | Received Blood Transfusion (n = 75) |
|---|---|---|
| Demographics | ||
| Age (Mean, SD) | 82, (7.66) | 85, (7.53) |
| Gender (M: F, % female) | 54 : 195, (78.3%) | 19 : 56, (74.7%) |
| Residence Pre-Injury n, (% of cohort) | ||
| Own Home | 184, (73.9%) | 57, (76%) |
| Residential Home | 42, (16.9%) | 11, (14.7%) |
| Nursing Home | 23, (9.2%) | 7, (9.3%) |
| ASA Grade n, (% of cohort) | ||
| 1 | 5, (2%) | 1, (1.3%) |
| 2 | 62, (24.9%) | 12, (16%) |
| 3 | 156, (62.7%) | 53, (70%) |
| 4 | 21, (8.4%) | 9, (12%) |
| 5 | 1, (0.4%) | 0, (0%) |
| Not Recorded | 4, (1.6%) | 0, (0%) |
| Anticoagulation Status at Presentation n, (% of cohort) | ||
| No Anticoagulation | 217, (87.1%) | 64, (85%) |
| Warfarin | 18, (7.2%) | 5, (6.6%) |
| Direct Oral Anticoagulant | 16, (6.4%) | 6, (8%) |
| Time to Surgery n, (% of cohort) | ||
| No Operation Performed | 22, (8.8%) | 0, (0%) |
| <36 h from presentation | 175, (70.2%) | 54, (72%) |
| >36 h from presentation | 52, (20.9%) | 21, (28%) |
3.2 Blood transfusions
75 patients (23%) received a PRC blood transfusion during their admission. 249 did not receive a transfusion. Three transfusions occurred pre-operatively, six intraoperatively, two in the theatre recovery area and 64 (85%) on the post-operative ward.
The mean admission haemoglobin (Hb) in the transfusion cohort was 110.1 g/L (SD 16.7). 28% of patients in the transfusion cohort had an admission Hb of less than 100 g/L (n = 21). The mean pre-transfusion Hb was 77.4 (SD 9.6) and the mean post-transfusion Hb was 94.6 (SD 9.3).
70.7% of transfused patients had a pre-transfusion Hb of less than 80 (n = 53). 94.6% of transfused patients had a pre-transfusion Hb of less than 90 (n = 70).
3.3 Anticoagulation
45 patients (13.9%) were taking an anticoagulant medication at presentation. 23 of these were taking Warfarin (51% of anticoagulated patients) and 22 a Direct Oral Anticoagulant (DOAC) (49% of anticoagulated patients). 27% of patients taking a DOAC (n = 6) and 21% of patients taking Warfarin (n = 5) required a transfusion.
24% of all anticoagulated patients (n = 11) required a blood transfusion. 14% of the transfusion cohort was anticoagulated at presentation. This is compared to 22.8% of patients who were not anticoagulated at presentation requiring a transfusion (n = 64).
There was no increased risk of transfusion if anticoagulant medication was being taken at presentation, odds ratio 1.097 (95% CI 0.5261 to 2.2873, p = 0.8050).
3.4 One year mortality
35.5% of patients presenting with hip fracture had died at one-year post injury (n = 115). In the transfusion cohort one-year mortality was 52% (n = 39), compared with 30.5% in the non-transfusion cohort (n = 76). The odds ratio of a patient who received a transfusion being dead at one-year post injury compared to a patient who did not receive a transfusion was 2.4660 (95% CI 1.4555 to 4.1780, p = 0.0008).
3.5 Binary logistic regression analysis
The odds ratio of a patient who received a transfusion being dead at one-year post injury compared to a patient who did not receive a transfusion was 2.790 (95% CI 1.554 to 5.010, p = 0.01). The model included age at presentation, ASA grade, sex, residence before admission, time in hours from presentation to surgery and if the patient received a transfusion.
4. Discussion
Many independent factors that can influence hip fracture mortality have been previously identified, such as early surgical intervention and the factors which make up the Nottingham Hip Fracture Score. Current National Institute for Health and Clinical Excellence (NICE) and British Orthopaedic Association (BOA) guidelines suggest early operative intervention improves outcomes.
7
,8
The Nottingham Hip Fracture Score is a validated score that can be used to predict 30 day and one-year mortality in hip fracture patients. It is made up of seven independent variables: age, sex (male), comorbidities, admission abbreviated mental test score (AMTS), pre-injury residence and the presence of malignancy.9
,10
Reversal of correctable preoperative abnormalities, such coagulopathy, and optimising perioperative medical management has significant benefits for these patients.16
,17
The management of perioperative anaemia is an essential element of surgical care. Studies have previously identified anaemia to be an independent risk factor for mortality, postoperative complication rates, length of hospital stay and poor functional outcomes. In the fractured neck of femur population anaemia can slow wound healing through reduced oxygen supply to tissues and hinder the post-operative rehabilitation of the patient. Anaemia may cause issues such as exertional breathlessness, dizziness on standing and lethargy which may all significantly affect the rehabilitation and mobilisation of the patient.
18
Whilst admission haemoglobin of less than 100 g/L has been shown to be predictive of mortality, we have not found it to be an accurate predictor of transfusion requirement. The mean admission haemoglobin of the transfusion cohort was 110.1 g/L, with just over a quarter presenting with a haemoglobin of less than 100 g/L. This shows admission haemoglobin to be a relatively poor predictor of transfusion requirement throughout admission. Most blood transfusions occurred in the postoperative setting, with only three transfusions occurring preoperatively. This suggests ongoing bleeding secondary to fracture and operative losses contribute more to transfusion requirement than baseline haemoglobin. Vigilance, early identification and prompt treatment of patients passing local thresholds for transfusion is therefore key.
We found almost a quarter (23%) of all hip fracture patients required a blood transfusion. Taking an anticoagulant medication at presentation did not increase a patient's risk of requiring blood transfusion. This suggests preoperative optimisation of these patients was effective at reducing operative losses, either through reversal of Vitamin K antagonists or allowing sufficient time between the last dose of Direct Oral Anticoagulants (DOACs) and surgery to allow for renal excretion.
19
, 20
, 21
Previous studies have shown both increased risk and no effect on mortality in hip fracture patients receiving a blood transfusion.
11
,22
, 23
, 24
, 25
Our study shows a statistically significant increased risk of mortality at one year if a patient received a blood transfusion compared to a patient who did not (OR 2.466, p < 0.05). As multiple previously described factors affect hip fracture mortality, binary logistic regression was performed to attempt to control for the effects of other predictors in a model. The model considered patient's age, sex, comorbidities (via ASA grade), residence before admission and time in hours from presentation to surgery. Admission haemoglobin was not included due to the high correlation between low haemoglobin and transfusion requirement.26
In our model, taking into account these other variables, receiving a transfusion still increased the risk of one-year mortality (OR 2.790, p < 0.05).There are multiple factors to consider when deciding whether to transfuse a patient. Low haemoglobin and comorbidities feature prominently in the internationally accepted practice of threshold-based transfusion.
13
, 14
, 15
At our organisation, a threshold of 80 g/L was set for the consideration of transfusion in hip fracture patients, or 90 g/L if cardiac comorbidities were present. 94.6% of transfusions met these thresholds. However, as with all interventions clinical judgment must play its part and consideration of individual patient circumstances should always be considered before a transfusion is administered.Within the transfused cohort, all transfusions were deemed clinically indicated by the prescribing physician. This decision was usually based on haemoglobin and comorbidities. Although we have observed a significantly increased mortality rate at one year in the transfused cohort, we do not know what the effect of not transfusing these patients would have been. If the transfusions were not given, would mortality rates be even higher at one year, or would they have remained the same with the same patients dying but perhaps earlier? PRC transfusion may delay mortality in this patient group in the short term, with no effect on their one-year mortality risk. If this is the case, receiving a PRC transfusion could be a useful way to identify patients who would benefit from closer follow up, in an attempt to mitigate this risk. The one-year mortality rate for all hip fracture patients presenting during the study period was 35.5%, which is in keeping with the national figure of approximately one third.
2
With an average admission haemoglobin of 110 g/L, most of our transfused cohort did not score a point for haemoglobin on the Nottingham hip fracture mortality prediction tool. When considering haemoglobin, the Nottingham score only looks at the admission result, with less than 100 g/L scoring on their system. We have observed that most of our transfused patients did not score on the Nottingham predictor tool for low haemoglobin on admission, however subsequently required transfusion and demonstrated a significantly higher mortality rate compared to the non-transfused cohort. Therefore, would a post op haemoglobin be a useful adjunct or addition to the Nottingham score in assessing a patient's individual risk of mortality at one year? If so, it could then be used to further counsel patients and their families regarding prognosis.
5. Conclusion
Hip fractures are common, serious injuries that carry significant morbidity and mortality for patients. Many factors have been shown to influence outcomes. Receiving a PRC blood transfusion during an admission for hip fracture carries an increased risk of one-year mortality of almost two and a half times. We suggest that the receipt of a PRC blood transfusion could be used to identify a high-risk subset of this already high-risk group, to aid in close follow up and potentially subsequent reductions in mortality.
Taking an anticoagulant medication at presentation did not increase the risk of requiring a PRC blood transfusion, highlighting the importance of proper preoperative optimisation in this comorbid patient group.
Only 28% of patients who received a transfusion had an admission haemoglobin of less than 100 g/L, showing it to be a poor predictor of PRC transfusion requirement.
94.6% of patients had a pre-transfusion haemoglobin of less than 90 g/L. This is in keeping with the internationally accepted practice of threshold-based transfusion. Early identification of patients requiring a PRC blood transfusion should be a goal in the management of hip fracture patients.
References
- Surgical interventions for treating extracapsular hip fractures in adults: a network meta-analysis.Cochrane library. 2019 Aug 19; : 2019
- National Hip Fracture Database Annual Report 2017.RCP London, 2016
- The changing epidemiology of fall-related fractures in adults.Injury. 2017 Apr; 48: 819-824
- Searching for the “winner” hip fracture patient: the effect of modifiable and non-modifiable factors on clinical outcomes following hip fracture surgery.Hip international. 2019 Sep 23; 31 (PMID: 31547719): 115-124
- Quality of life after hip fracture in the elderly: a systematic literature review.Injury. 2016; 47: 1369-1382
- Predictive factors associated with mortality in Korean elderly patients with hip fractures.J Orthop Surg. 2019 May 31; 27 (2309499019847848)
- Patients Sustaining a Fragility Hip Fracture.2012 January
- Hip fracture management.Clinical Guideline (CG124). 2017 May; : 5
- Prediction of 30-day mortality after hip fracture surgery by the Nottingham hip fracture score and the surgical outcome risk tool.Anaesthesia. 2016 May; 71: 515-521
- Validation of the Nottingham Hip Fracture Score (NHFS) to predict 30-day mortality in patients with an intracapsular hip fracture.Orthopaedics & traumatology, surgery & research. 2019 May; 105: 485-489
- The risks of red cell transfusion for hip fracture surgery in the elderly.Vox Sang. 2012 Oct; 103: 223-230
- Body mass index, risk of allogeneic red blood cell transfusion, and mortality in elderly patients undergoing hip fracture surgery.Osteoporos Int. 2016 Sep; 27: 2765-2775
- Blood Transfusion (NG24).2015 November
- Joint United Kingdom (UK) blood transfusion and tissue transplantation services professional advisory committee.in: Handbook of Transfusion Medicine. 5th Edition. Publisher: TSO information and publishing solutions, 2014 January: 75
- Red blood cell transfusion: a clinical practice guideline from the AABB.Ann Intern Med. 2012 Jul 3; 157: 49-58
- Management of proximal femoral fractures 2011: association of anaesthetists of great britain and Ireland.Anaesthesia. 2012 Jan; 67: 85-98
- Evolution of the hip fracture population: time to consider the future? A retrospective observational analysis.BMJ open. 2014 Apr; 4 (e004405)
- Anaemia in the older surgical patient: a review of prevalence, causes, implications and management.J R Soc Med. 2013 Jun 6; 106: 269-277
- Management of novel oral anticoagulants in emergency and trauma surgery.Surgeon. 2016; 14: 234-239
- Use of anticoagulants remains a significant threat to timely hip fracture surgery.Geriatric orthopaedic surgery & rehabilitation. 2018 Mar 22; 9 (215145931876415-2151459318764150)
- Venous thromboembolism in 2013: the advent of the novel oral anticoagulants.Nat Rev Cardiol. 2014 Feb; 11: 70-72
- Blood transfusion rates and predictors following geriatric hip fracture surgery.Hip international. 2020 Jan 8; 31 (PMID: 31912747): 272-279
- Effect of blood transfusion on survival after hip fracture surgery.Eur J Orthop Surg Traumatol. 2018 May 11; 28: 1297-1303
- Risk factors and mortality of patients undergoing hip fracture surgery: a one-year follow-up study.Sci Rep. 2020 Dec; 10: 9607
- A systematic review of pre-operative anaemia and blood transfusion in patients with fractured hips.Anaesthesia. 2015 Apr; 70: 483-500
- Common pitfalls in statistical analysis: logistic regression.Perspectives in clinical research. 2017 Jul; 8: 148-151
Article info
Publication history
Published online: July 19, 2021
Accepted:
July 13,
2021
Received in revised form:
July 12,
2021
Received:
June 20,
2021
Identification
Copyright
© 2021 Delhi Orthopedic Association. All rights reserved.
