Key words
anaemia - blood loss - blood transfusion - orthopaedic surgery
Introduction
A haemoglobin concentration of < 12 g/dl in women and < 13 g/dl in men is defined
as anaemia according to the World Health Organisation (WHO) with an increased need
for transfusion of allogeneic red blood cell (RBC) concentrates. Baron and colleagues
[1] recently analysed about 40,000 surgical patients from 28 European countries. Surprisingly,
almost one in three patients showed anaemia associated with prolonged hospital stay
and increased risk of hospital mortality prior to surgery. In the PREPARE study by
Lasocki and colleagues [2], a total of 1,534 patients were evaluated in 17 European centres for elective knee
and hip joint endoprosthetics and spinal surgery. The prevalence of anaemia before
surgery was 14.1% and even increased to over 85.0% at the time of hospital discharge.
In the meantime, numerous studies have proven the effectiveness of pre-operative anaemia
management [3], [4], [5], [6], [7], [8], [9]. Likewise, the directive ‘Hemotherapy 2017’ of the German Medical Association [10] requires that ‘prior to substitution treatment with blood products, patients must
be examined individually on the basis of current findings whether other measures are
suitable to remedy chronic or acute deficiencies. These include optimising the volume
of red blood cells, minimising bleeding and blood loss as well as increasing and fully
exploiting anaemia tolerance (Patient Blood Management)’.
In this study the incidence of pre-operative anaemia and its influence on the transfusion
needs of RBC, hospital stay and hospital mortality in primary hip and knee joint endoprosthetics
shall be analysed.
Patients and Methods
Anonymous register data from 13 hospitals of the accompanying epidemiological research
project of the German PBM Network [11] were used as a data basis [11]; the protocol was approved by the respective local ethics committees (lead ethics
committee: University Hospital Frankfurt 380/12). Data of adult inpatients (over 18
years of age) undergoing surgery were recorded (ClinicalTrials.gov Identifier: NCT02147795).
In addition, a data protection vote of the Hessian data protection officer has been
obtained (Ref: 43.60; 60.01.21-ga). The data were extracted from the respective electronic
hospital information systems and anonymised for further analysis. Patient-specific
observation period ranged from hospital admission to discharge.
Patient Population
Patients with hip and knee joint primary implantation in gonarthrosis, coxarthrosis
and femoral neck fracture coded with the following Operationen- und Prozedurenschlüssel
(OPS) codes (as of June 2019) were included in the analysis. No exclusion criteria
were defined.
-
Knee joint (elective primary implantation in gonarthrosis): 5-822.00-02, 5-822.83-87,
5-822.g0-g2, 5-822.j0-j2, 5-822.k0-k2.
-
Hip joint (elective primary implantation in coxarthrosis): 5-820.00-02, 5-820.30-31,
5-820.80-82, 5-820.92-96.
-
Hip joint (dual head prosthesis in femoral neck fracture): 5-820.40-41.
Endpoints
The primary endpoint was the incidence of pre-operative anaemia, which was analysed
by the first available pre-operative haemoglobin value according to the WHO definition.
Secondary endpoints included hospital stay, number of patients with RBC transfusions,
number of transfused RBC per 1000 patients, incidence of post-operative hospital-acquired
anaemia, number of patients deceased in hospital (death of any cause as discharge
reason), number of patients with acute renal failure (ICD-10: N17.0, N17.1, N17.2,
N17.8, N17.9, N19, N99.0) and pneumonia (viral pneumonia (J12.0 – J12.3, J12.8, J12.9),
pneumonia caused by Streptococcus pneumonia (J13), pneumonia caused by Haemophilus influenzae (J14), pneumonia caused by bacteria (J15.0 – J15.9), pneumonia caused by other infectious
agents (J16.0, J16.8), pneumonia, pathogen not specified (J18.0 – J18.2, J18.8, J18.9)).
The results were also broken down by age (< 80 and ≥ 80 years).
Statistical analysis
The statistical evaluation included for all endpoints a detailed descriptive analysis
broken down by three surgical groups and two age groups (mean values ± standard error,
medians with first and third quartile, percentage rates with 95% confidence interval).
Furthermore, non-parametric significance tests (Mann–Whitney or Fisherʼs exact test,
p values) for differences between the relevant subgroups (with/without RBC transfusion
or pre-operative anaemia) were performed for the endpoints mortality and hospital
stay as well as for the estimation of the magnitude of RBC transfusion or pre-operative
anaemia (Spearmanʼs rank correlation coefficient, r).
Finally, to investigate the dependence of the endpoints mortality and hospital stay
on the influencing factors RBC transfusion and pre-operative anaemia, an evaluation
was performed using a multi-variate mixed regression model which included the hospitals
as random effects to calculate the hospital-specific cluster effects as well as RBC
transfusion and pre-operative anaemia as the fixed effects in addition to age, gender
and surgery group (if significant) as influencing variables. Potential influence variables
for each model were set a priori and then tested with univariate analysis (non-parametric and in the regression model)
for significance before being included in the final model.
Results
From January 2012 to September 2018, routine data from a total of 378,069 patients
from 13 German hospitals were analysed, of which 10,017 patients underwent hip and
knee joint primary endoprosthetics. One patient was excluded from the analysis because
he was marked as both an elective hip joint and dual head prosthesis.
The incidence of pre-operative anaemia was 14.8% in elective knee joint endoprosthetics,
22.9% in elective hip joint endoprosthetics and 45.0% in implantation of a dual head
prosthesis ([Tab. 1]). Pre-operative anaemia resulted in a significantly higher rate of RBC transfusion
(32.7% vs. 7.3%, Fisher p < 0.001) (knee joint prosthesis): 8.3% vs. 1.8%, p < 0.001; hip joint prosthesis: 34.5% vs. 8.1%, p < 0.001; dual head prosthesis: 42.3% vs. 17.4%, p < 0.001) and RBC consumption (989 ± 50 vs. 174 ± 11 RBC/1000 patients, Mann–Whitney p < 0.001) (knee joint: 256 ± 107 vs. 29 ± 5, p < 0.001; hip joint: 929 ± 60 vs. 190 ± 16, p < 0.001; dual head prosthesis: 1411 ± 98 vs. 453 ± 42, p < 0.001). In multivariate regression analysis, rates of RBC transfusion
and RBC consumption were also significantly higher in the presence of pre-operative
anaemia (p < 0.001).
Table 1 Demographic data, anaemia rate, transfusion rate, RBC consumption and clinical endpoints.
|
Primary implantation
|
Knee prosthesis (n = 3162)
|
Hip prosthesis (n = 4813)
|
Dual head prosthesis (n = 2041)
|
|
* n = 1251 (12.5%) no pre-operative Hb values; ** n = 1163 (11.6%) no post-operative
Hb values; # p < 0.001 no anaemia vs. anaemia ( Fisher), p < 0.001 in all individual surgical groups and in the total cohort;
## Mann–Whitney p < 0.001 in all individual surgical groups and in the total cohort.
Hb = haemoglobin; WHO = World Health Organization
|
|
Age, years
|
70.0 [62.0 – 76.0]
|
71.0 [62.0 – 78.0]
|
84.0 [78.0; 89.0]
|
|
Gender, male, %
|
39.7
|
44.2
|
33.2
|
|
Hb pre-operative,* g/dl
|
13.7 [12.8 – 14.7]
|
13.5 [12.4; 14.5]
|
12.4 [11.1; 13.6]
|
|
Pre-operative anaemia (WHO), % (n)*
|
14.8 [13.5 – 16.2] (398)
|
22.9 [21.7 – 24.2] (974)
|
45.0 [42.7 – 47.3] (822)
|
|
Hb post-operative,** g/dl
|
11.0 [10.0; 12.1]
|
10.2 [9.3 – 11.2]
|
10.0 [9.0 – 10.9]
|
|
Hospital-acquired anaemia (WHO), % (n)*
|
81.2 [79.6 – 82.6] (2184)
|
93.6 [92.8 – 94.3] (3996)
|
93.2% [92.0 – 94.3] (1764)
|
|
Length of hospital stay, days
|
10.0 [9.0 – 12.0]
|
11.0 [9.0 – 13.0]
|
13.0 [10.0 – 18.0]
|
|
Hospital mortality, % [95% CI] (n)
|
0.1 [0.0 – 0.2] (2)
|
0.8 [0.6%–1.1] (39)
|
7.4 [6.3 – 8.7] (152)
|
|
Pneumonia, % [95% CI] (n)
|
0.4 [0.2 – 0.7] (14)
|
1.7% [1.4 – 2.1] (83)
|
10.5 [9.2 – 11.9] (214)
|
|
Renal failure, % [95% CI] (n)
|
3.1% [2.5 – 3.8] (98)
|
4.7% [4.2 – 5.4] (228)
|
14.3 [12.8 – 15.9] (291)
|
|
RBC transfusion rate, % [95% CI] (n)
|
|
|
|
|
No Pre-operative anaemia
|
1.8 [1.3 – 2.5] (42)
|
8.1 [7.2 – 9.1] (265)
|
17.4 [15.1 – 19.9] (175)
|
|
Pre-operative anaemia
|
8.3 [5.8 – 11.4]# (33)
|
34.5% [31.5 – 37.6]# (336)
|
42.3 [38.9 – 45.8]# (348)
|
|
RBC/1000 patients; mean ± standard error
|
|
|
|
|
No pre-operative anaemia
|
29 ± 5
|
190 ± 16
|
453 ± 42
|
|
Pre-operative anaemia
|
256 ± 107##
|
929 ± 60##
|
1411 ± 98##
|
Compared with non-anaemic patients, those with pre-operative anaemia showed significantly
prolonged hospital stay [hospital stay: 12.0 (10.0 – 17.0) days vs. 11.0 (9.0 – 13.0) days; Mann–Whitney and multi-variate regression model p < 0.001]
and increased mortality [5.5% (4.6 – 6.5%) vs. 0.9% (0.7 – 1.2%); Fisher and multi-variate regression model p < 0.001]. Likewise,
the transfusion of at least one RBC was significantly correlated to prolonged hospital
stay [15.0 (11.0 – 22.0) days vs. 11.0 (9.0 – 13.0) days; Mann–Whitney and multi-variate regression model p < 0.001]
and increased mortality rate [7.4% (6.1 – 9.0%) vs. 1.1% (0.9 – 1.4%); Fisher and multi-variant regression model p < 0.001].
Patients aged 80 years and older showed a significantly higher overall RBC transfusion
rate (24.2% vs. 8.7%, Fisher p < 0.001), higher RBC consumption (601 ± 29 vs. 256 ± 152 RBC/1000 patients; Mann–Whitney p < 0.001), longer hospital stay (14.6 ± 0.2
vs. 12.1 ± 0.1 days; Mann–Whitney p < 0.001) and higher mortality rate (5.1% vs. 0.8%; Fisher p < 0.001) than patients under 80 years of age ([Fig. 1]).
Fig. 1 Transfusion rate of red blood cell (RBC) concentrates in patients without (blue)
and with anaemia (orange) in elective knee and hip joint endoprosthetics as well as
implantation of a dual head prosthesis at the age of < 80 or and ≥ 80 years.
Discussion
In about every third patient, untreated pre-operative anaemia can be observed [1], [12]. Various observational studies have concluded that pre-operative anaemia must be
considered an independent risk factor for RBC transfusion, potential complications
and post-operative mortality [1], [13], [14]. In our analysis, the prevalence of pre-operative anaemia was 14.8% for elective
knee joint endoprosthetics, 22.9% for elective hip joint endoprosthetics and 45.0%
for dual head prosthesis implantation. In the literature, different anaemia definitions
are frequently used and the surgery on the hip or knee are only analysed collectively.
A meta-analysis of 19 studies including 35,000 patients showed anaemia and transfusion
rates of 24% ± 9% and 45% ± 25%, respectively, for elective hip and knee endoprosthetics
and 51% and 44% ± 15%, respectively, for hip fracture [14]. In our analysis, pre-operative anaemia was associated with up to 10-fold increased
RBC consumption, whereas other studies have shown 4 – 11-fold increased probability
for RBC transfusions [2], [15], [16], [17].
In this context, there is great potential for pre-operative anaemia management, as
required by various international guidelines [18], [19], [20], by the current S3 guideline ‘Preoperative Anemia’ [21] and by the directive ‘Hemotherapy 2017’ [10].
The causes of anaemia in orthopaedic/trauma patients are multifaceted [6], [22], [23], [24]. At least in a subgroup of these patients, anaemia is caused by an iron deficiency.
In a recent Danish study of orthopaedic knee and hip patients, Jans et al. reported
that > 40% of the anaemic patients presented with iron deficiency [15]. Possible causes of iron deficiency include malnutrition (vegetarians, vegans or
alcohol addicts); dysfunctional enteral iron absorption (e.g. Helicobacter pylori gastritis, gastrectomy, atrophic gastritis, chronic inflammatory bowel disease, medication
with NSAIDs or proton pump inhibitors) or chronic blood loss (e.g. angiodysplasia,
neoplasia, ulcers, diverticula, or increased menstruation).
Based on the current S3 guideline [21], anaemia diagnosis should be initiated preoperatively in a timely manner. In this
respect, early identification (at least 2 – 4 weeks preoperatively) of anaemic patients
is crucial in the context of preparations for surgery. If iron deficiency has been
diagnosed, therapy should be started primarily with iron based on the cause. However,
quite a few patients discontinue oral iron substitution due to gastrointestinal problems
and intolerances [25]. If oral iron therapy is ineffective or unsuitable, such as in the case of urgency
of the procedure (< 6 weeks), parenteral iron therapy is recommended in principle,
whereby the available preparations differ from each other, particular in their complex
stability and safety [5], [26], [27], [28]. In case of chronic anaemia or renal anaemia (erythropoietin deficiency), treatment
with erythropoietin alone or, in case of additional iron deficiency, in combination
with iron is recommended [21]. Theusinger et al. [16], for instance, have confirmed the benefit of pre-operative anaemia management, particularly
in patients undergoing orthopaedic elective procedures. In elective knee joint endoprosthetics,
the anaemia rate was reduced from 15.5% to 7.8% and the transfusion rate from 19.3%
to 4.9%. In elective hip joint endoprosthetics the anaemia rate was reduced from 17.6%
to 12.9% and the transfusion rate from 21.8% to 15.7% [16]. Kotze et al. have found similar achievements in pre-operative anaemia management
in orthopaedic patients [17]. This study now focusses on the demographic change of the geriatric orthopaedic
patient population. Our analysis shows that in the field of elective orthopaedic patient
care, the pre-operative anaemia rate in patients over 80 years of age is significantly
higher than that in patients under 80 years of age (knee prosthesis: 23.9% vs. 13.4%; hip prosthesis: 37.2% vs. 19.7%). These data support the need to diagnose and treat pre-operative anaemia as
part of future geriatric orthopaedic concepts in elective surgery.
Unlike elective patients, patients with an acute femoral fracture who received dual
head prosthesis in an emergency within 24 – 48 hours pose a particular challenge.
In these cases, anaemia rate was as high as 45% and EC transfusion rate over 40%.
Since pre-operative time is very limited, new concepts for perioperative anaemia management
are warranted. A simple laboratory diagnostic using ferritin on the day of surgery
would be conceivable here, for example, in order to be able to conduct parenteral
substitution with iron indicated at the end of the surgery or postoperatively, particularly
in cases of perioperative blood loss of > 500 ml [29], [30]. At least in the field of cardiac surgery, it has recently been demonstrated that
an ultra-short term combined administration of intravenous iron, erythropoietin alpha,
vitamin B12 and folic acid could reduce post-operative RBC consumption by 1 day preoperatively
[31].
A substantial advantage of the current analysis is the comprehensive representative
analysis of routine data of more than 10,000 patients with primary hip and knee joint
endoprosthetics at 13 German hospitals. Anonymous data from an established clinical
patient register were used as the data basis. Clinical registries are increasingly
being used in healthcare to gain knowledge. Existing data sources can thus be used
and evaluated comprehensively. However, the current register analysis has the limitation
that unknown factors for RBC transfusion or post-operative complications could not
be ruled out (e.g. use of ischaemia, tranexamic acid, duration of drainage stay or
wound and joint infections) although risk adjustments have been performed. In some
patients, pre-operative anaemia management may even have been performed locally as
part of patient blood management. However, since pre-operative anaemia management
cannot be coded based on either OPS or the International Classification of Diseases,
the potential influence of any anaemia management could not be taken into account.
Conclusion
In summary, the current analysis reported pre-operative anaemia rate of 14.8% in elective
knee joint endoprosthetics, 22.9% in elective hip joint endoprosthetics and as high
as 45.0% in (emergency) dual head prosthesis implantation. Pre-operative anaemia led
to a relevant increase in RBC consumption and was associated with prolonged hospital
stay and increased mortality. In patients aged 80 years and older, the incidence of
pre-operative anaemia and consequent transfusion rate were almost two times higher
than in patients under 80 years of age. In this context, a relevant potential could
arise in the future, particularly in elective geriatric orthopaedics, to better prepare
elective patients in terms of patient blood management, to avoid unnecessary RBC transfusions
and to thus conserve the valuable resource blood [32].
