zyxw
zyxwv
zyxwvutsrqpo
z
State-of-the-Art Paper
Vox Sanguinis
L. Noel
J. Debeir
A. Cosson
Vox Sanguinis 1998; 74 (Suppl. 2): 441-445
Received March, 1998
Accepted: May, 1998
The French Haemovigilance System
zyxwvu
zyxwvutsr
...............................................................................................
Agence Franqaise du Sang
Paris
France
Abstract
Haemovigilance was part of the reform of the French transfusion system. The
haemovigilance network is now operational with approximately 4600
transfusion incidents reported annually. Immediate incidents observed within 8
days after transfusion account for 85 % of the reports. A cause cannot be
identified in 41% of these, usually concerning minor clinical incidents with
transient fever and /or shivers. An allergic reaction is described in 31% of
transfusion reactions. Immunological conflicts account for 18% and bacteria
associated transfusion reactions for 6%. The importance of bacteria associated
transfusion reactions, the first identified cause of death associated with
transfusion is one of the findings of haemovigilance. Improvement in the
haemovigilance systems aims at obtaining better descriptions of transfusion
incidents, standardisation of severity and imputability assessment and
definitions of denominators such as the actual number of recipients. Delayed
incidents will ultimately provide a true vision of post transfusion immunisation
and infection The improvement of haemovigilance now considered as part of
transfusion medicine practice is a continuous process.
....................
In 1992 the French health authorities undertook a
thorough reorganisation of the transfusion system in
reaction to the trauma of the tainted blood affair and at a
time when transfusion appeared as one of the major routes
of HCV transmission. One of the innovative features of
the Blood Transfusion Safety Act of 1993 was the concept
of haemovigilance. Haemovigilance is a national system of
surveillance and alarm, from blood collection to the
follow-up of the recipients. The d g e n c e Franqaise du
Sang)) (AFS) was entrusted to set forth the
haemovigilance system, starting in 1994 .
The aim of Haemovigilance is to detect, gather ad
analyse all untoward effects of blood transfusion in order
to correct their cause and prevent recurrence. Causes of
transfusion untoward effects may be found at all levels,
from donor selection to the transfusion act. Therefore the
scope of haemovigilance was defined as encompassing all
steps of the transfusion process, up to recipient follow-up.
01998 S. Karger AG, Basel
0042-9007/98/0746-044 1 S15 .OO/O
1. Concept and organisation of haemovigilance
zyx
In the context of haemovigilance inception there was a
need for prompt efficacy: a reliable haemovigilance system
was deemed necessary to restore confidence in the
transfusion service and react adequately to future
problems. The means then chosen were adapted to the
existing situation in France.
A. A national system created by law
The choice of a national haemovigilance system driven
by a government agency provided the authority needed to
federate all actors whether in hospital or in blood centres
and warrant exhaustivity. Llkewise a legal obligation to all
health care personnel to report any unexpected or
undesirable effect associated to transfusion was considered
a prerequisite to reach exhaustivity. This was also a sign
of the new importance attached to transfusion.
L. Noel
Agence Franqaise du Sang
6 rue Alexandre Cabanel
75015 Pans, France
B. A dedicated haemovigilance network
zyxwvu
zyxwvu
To a network of haemovigilance correspondents, usually
physicians, one for every blood centre and every hospital
performing transfusion, was entrusted the responsibility to
collect reports on transfusion adverse events and relay
them to the local health authorities and to the A F S
centralised haemovigilance cell. Full time regional coordinators supervise the application of haemovigilance.
Such a heavy organisation was needed to compensate
for the lack of hospital blood banks or proper transfusion
service in more than 90 per cent of the hospitals or clinics
performing transfusion. Indeed in France, the
immunological security of red cells transfusions always
has been based on type and screen. Cross matching is far
from systematic in most centres. Thus blood products can
be ordered from the blood centres or taken from a deposit
on the basis of tests results from any clinical laboratory.
This accounted for a poor traceability and a lack of
expertise in transfusion medicine in many hospitals.
information from the hospital to the blood centre
confirming the actual recipient’s identity.
Similarly, transfusion incidents reports have to be cosigned by both haemovigilance correspondents, from the
hospital and from the blood centre.
E. Management of transfusion incident reports
Reports on transfusion incidents (TIR) are centralised at
the AFS level. The AFS haemovigilance cell has to look
out for incidents that could involve multiple recipients or
reveal a potential serial risk. This early warning function
must allow a prompt reaction. In addition “cold” analysis
of transfusion incidents aims at revealing trends and
provides a better understanding of transfusion morbidity.
The AFS also initiates and funds prospective studies in
haemovigilance, in particular those carried out by the
(( Centre National
d’HCmovigilanceB, from Bordeaux
university.
Haemovigilance results are communicated to the
medical community through presentations at various
scientific meetings and regularly updated in the (( Bulletin
d’HCmovigilance>> published by the AFS.
zy
zyxwvuts
zyxwv
C. Recognition of transfusion as a hospital
activity
The lack of a centralised transfusion expertise in most
hospitals had to be compensated by the mandatory
creation of transfusion security and haemovigilance
committees H in public hospitals and the designation of
haemovigilance correspondents in all health care
institutions performing transfusions.
Training of these correspondents is carried out by the
(( Institut National de Transfusion Sanguine )> (INTS) in
connection with the AFS. Since 1994, 650 haemovigilance
correspondents followed one of the dedicated training
sessions at the INTS.
D. Definition of formal relationships between
Hospital and blood centres
Haemovigilance aims at improving links between
hospitals and blood centres. A given hospital can only
have one blood centre as a provider of blood products.
Traceability is a constituent of hEmovigilance, defined as
the capacity to identify the recipient for any given blood
component and conversely to be able to trace all
components received by a given patient. The requirements
for a thorough traceability of blood products involve a
legal obligation for the blood centres as well as the
hospital to keep a register of the recipient‘s identity for
each blood product. This requires a post transfusion
442
II. Achievements of haemovigilance
A. The French hamovigilance network is
operational
Nearly four years after its founding, haemovigilance has
reached its cruising speed. The network is completed.
Some 2000 correspondents, co-ordinated by 24 regional
co-ordinators, send transfusion incident reports that are
ultimately analysed at the AFS haemovigilance cell. In the
latter three physicians are involved ( two specialised in
transfusion medicine, one in public health), and an
epidemiologist and a statistician.
The haemovigilance network has been equipped with an
electronic mailing system and specific software, allowing
TIR transfer from the blood centres to the regional coordinators and to the haemovigilance cell as well as e-mail
exchanges and broadcasting of informations.
B. Analysis of TlRs
Since 1994 to October 1st 1997 10880 completed TIR
have been registered.. The system is now in routine with
4600 to 4900 TIRs reported annually. The rate of reported
zyxwvutsrqpo
Vox Sang 1998; 74 (Suppl2): 441-445
Noel/Debeir/Cosson
transfusion incident per delivered blood product is
1.5/1000.
The delay before notification of transfusion incidents
shows the performance of the haemovigilance network. In
46% of cases TIR are notified within one day, in 23% of
cases on day 2 and in 20% of incidents notification takes
place between days 3 to 5 .
Among all TIRs that could be analysed, 85.2% describe
an immediate incident, a reaction occurring within 8 days
of transfusion. In 4.4% a delayed incident and in 11.2% a
post transfusion red cell immunisation.
The other main causes of transfusion reactions were
immunological conflict (18%) and bacteria associated
transfusion reactions (6%) .
zyxwvuts
zyxwvutsrqpo
zyxwvut
zyxwvu
zyxwvut
zyxwvutsr
zyxwvutsrqp
Analysis on TIRs is carried out according to severity
and imputability:
0
The severity of a transfusion incident is graded by the
correspondents from 1 to 4. Grade 1, accounting for
82% of reports, describes minor symptoms; grade 2
applies to incidents with long term consequences ( 13
% of TIRs); grade 3 describes a vital threat as reported
in 4 % of TIRs and grade 4 the death of the recipient,
reported in 0.73% of TIRs.
Imputability is also scored by the correspondents in the
field. In 19.30 % of TIRs, the role of transfusion is
excluded or dubious ; in 32.29 YOof TIRs, transfusion
may have played a role in the incident ; in 31% its
responsibility is probable and in 7.5% of cases
transfusion definitely was the cause of the incident. For
example, in 33 out of 79 cases of death reported as
transfusion incidents, the responsibility of transfusion
is either considered excluded (n=15) or dubious
(n=18).
It should be noted that the severity scale does not take
into account the patient’s pre-transfusion state.
Transfusion of an unstable patient can be a cause of death
or vital threat when the same reaction would have been
graded as minor in a less vulnerable recipient. This is
probably one of the reasons why incidents of unknown
cause rank first among the cases of death associated with
transfusion.
I . Immediate incidents
a. Causes of transfusion reactions
Analysis of causes was carried out on 6767 TIR
selected for an imputability graded from possible to
unquestionable. In 41% of early transfusion incidents,
observed within 8 days of transfusion, a cause has not
been identified. In the vast majority of these cases
symptoms essentially consisted in fever andor shivers
with no identified immunological conflict. The role of
cytokines, whether active or passive, remains to be
properly investigated.
In 31% of TIR anaphylactoid symptoms suggested an
allergic reaction.
The French Haemovigilance System
b. HaemoIytic transfusion reactions
ABO mismatch has been reported with allogenic red
cell concentrates in 58 cases since the beginning of
haemovigilance, in 4 cases resulting in the death of the
recipient (approximately 1 per 2 million units). More than
one ABO mismatch is reported per month. despite the fact
that a pretransfusion bed side check of ABO compatibility
is required by law in France. It should be noted that
several cases of ABO mismatch with no clinical
consequences, not part of haemovigilance strict0 sensu ,
have been reported, a fact that strengthens confidence in
the exhaustivity of haemovigilance.
Non-ABO haemolytic transfusion reactions reported in
117 TIRs, predominantly involve antibody against
antigens of the RH (39), JK (3 1) and FY (13) blood group
systems. Death occurred in 4 cases involving anti-JK1,
anti-JK2, anti-KEL1 and anti-MNS3.
c. Bacteria associated transfusion reactions
Bacteria associated transfusion reactions appeared as
the most frequently by identified cause of death in
transfusion incidents. Since the beginning of
hEmovigilance, 16 deaths have been attributed to bacteria
infected blood products, 4 of them in 1997. The
importance of bacteria in transfusion reactions raises a
number of issues regarding methods of investigation,
sources of contamination of blood components, possibility
of transfusions affecting a previously unrecognised
infection focus in the recipient. The concept of (< bacteria
associated transfusion reactions >> covers transfusion
incidents involving bacteria and includes bacteria infected
blood components. Recommendations on prevention and
investigations were issued to blood centres and hospitals.
A prospective case control study of all bacteria associated
transfusion reactions has been ongoing since the end of
1996.
d. Hcemovigilance provides an overall vision of
transfusion reactions
Results of haemovigilance increased awareness to
known transfusion risks. For example several cases of
deaths by vascular overload associated to transfusion of
red cell concentrates prompted to remind the clinicians of
this apparently trivial risk.
On the other end haemovigilance revealed less classical
potentially by adverse effects of transfusion: for example a
study group is investigating cases of post transfusion
necrotising enterocolitis in the new born.
The haemovigilance cell did not identify any cases of
potentially serial incidents associated with a defective
disposable since 1994. After the recent description of
Vox Sang 1998; 74 (Suppl2) : 441-445
443
zyxw
zyxwvutsrq
and their colleagues. Regular exchanges, feed back of
informations, formal training via the regional
hamovigilance co-ordinators are essential. Closely
working with the hamovigilance cell, the group of
regional co-ordinators succeeded in reducing regional
efficacy variations.
A danger for hamovigilance would be for it to be
perceived as repressive. Care have to be given to make
2. Delayed transfusion incidents
sure that it remains descriptive and constructive.
Most delayed transfusion incidents are immunisations
In fact hamovigilance is one of a new type of
to red cell antigens (971 TIR).
responsibility in a hospital, associated to the other
At the present time data on viral transmission since vigilances, drugs or medical devices for instance.
1994 are scarce: 2 cases of HIV transmission by However hamovigilance usually relies on one of the
transfusion have been reported, in both cases the unit physicians. Progress in medical evaluation and the search
involved was antibody and p24 antigen negative but for quality assurance in health care will ultimately bring
positive by PCR according to the repository samples, only vigilance functions to be a natural component of medical
one HCV and one HBV transmission have been properly practice and therefore of hospital teams.
documented since 1994.
The improvement of haemovigilance is a continuous
Systematic pre and post transfusion testing to process. Transfusion incident report forms are in the
demonstrate red cell immunisation or viral transmission as process of being modified according to the lessons of
recommended by the French health authorities in October experience. Standardisation is the main issue, whether of
the clinical description, of the investigation of the
1996 will improve the detection of these risks.
potential causes or in imputability assessment.
No cases of parasite transmission has been reported.
Nosology of transfusion reactions needs to be improved
3. Risks associated with pre-deposit autologous as judged by a majority of incidents of unknown cause.
For example transfusion related acute lung injuries are not
transfusion
always well differentiated from vascular overload in their
In 1996 6,5% of red cell concentrates issued in France minor forms. Likewise what would be the criteria to
were autologous. Among transfusion incidents reports, 1% ascribe subsequent bacterial infections to transfusion
involve autologous blood products. Most of these are related immune-modulation?
benign transient transfusion reactions. However recipient
On the other hand hamovigilance needs to remain open
mismatch with a autologous D red cell concentrates is to the unexpected. Any incident remotely associated to
reported in 5 cases, with a severe ABO conflict in 2 cases. transfusion has to be reported, it may ultimately be part of
Bacterial infection of autologous red cell concentrates a pattern. Consequently analysis of the hamovigilance
resulted in 3 cases of vital threat, 2 with Y. enterocolitica, data base has to stratify TIRs according to imputability.
1 with E. cloacae.
The role of the AFS hamovigilance cell is particularly
important. As its converging centre, it animates the
network and identifies or endorses reports requiring
further action. In this respect, as the government agency
111. Limitations of haemovigilance
overseeing transfusion, the AFS has the responsibility to
take any corrective measure deemed necessary and coAlthough haemovigilance is nearly 4 years old, data ordinate reactions.
concerning recipients and actual blood usage are still not
available while they would be the needed as denominators
of hamovigilance results. One of the reasons lies in the
IV. Perspectives
heterogeneity of the informations systems used by blood
centres and hospitals in France. However the AFS has
developed communication standards to facilitate data
An important effort resulted in the creation of
exchanges within the transfusion service and between haemovigilance. The next step is to optimise its operation
blood centres and hospitals. These are progressively and strengthen the quality of collected data. Clinical and
implemented and will result in improving data. .
biological investigations of adverse effects of transfusion
The haemovigilance network is still fragile: hospital will have to be improved by taking advantage of the
correspondents often have been charged with national system to organise prospective studies and
haemovigilance as an additional responsibility without any centralise testing whenever necessary.
compensation. They need the support of their institution
adverse ocular reactions following transfusion in the USA
(MMWR,1998, 47,3 :49-50), a query to the
haemovigilance data base did not reveal a single similar
case indirectly confirming the association of this incident
with a brand of leucodepletion filters not marketed in
France.
zyxwvu
zyxwvut
zyxwvutsrqp
zyxwvuts
zyxwvuts
444
Vox Sang 1998; 74 (Suppl2): 441-445
Noel/Debeir/Cosson
The specific software used to transfer and analyse TIR
will be improved to provide more controls and detect
anomalies on pre-established criteria.
Communication with all actors in particular with
clinicians will have to be amplified favouring informations
exchanges and availability of reference protocols, using
Internet for easy access.
A further evolution of the French transfusion service
will soon take place. The law presently under discussion
allows for a single operator of the French transfusion
service, the
Etablissement FranGais du Sang (( (ESF),
with the present centres as its subsidiaries. The
hamovigilance cell will have to be associated to the EFS
to preserve the capacity needed for investigations and
reaction. It will then be in a better position to homogenise
the centres haemovigilance activities. Conversely the new
Sanitary Agency will oversee the transfusion service and
will have to be involved in haemovigilance. It will provide
the needed links with other public health surveillance
systems.
There are three prerequisites for the success of
hamovigilance: 1) proximity to the blood centres,
allowing for an adequate reaction with the minimum
delay; 2) independence from the actors involved in the
various links of the transfusion chain and 3) support of the
health authorities granting access to the clinical teams.
Haemovigilance watches over the failures of transfusion
therapy to improve its practice. The clue to a successful
haemovigilance program relies on maintaining an efficient
cooperations between blood centres and hospitals.
Development of hospital based transfusion services will
have to take place. The role of the EFS may vary but in all
cases a strong link of the EFS with the transfusion services
will be needed.
A full-blown transfusion medicine service should
include, as part of its routine process, an evaluation of the
result of each transfusion. Haemovigilance ultimately will
be integrated in the quality assurance program. Legal
obligation is not enough to maintain and improve an
efficient haemovigilance system, haemovigilance should be
understood as a tool of medical practice.
zy
zyxwvut
zyxwvu
V. Conclusions
The vulnerability of blood and the potential risks of
transfusion justified the specificity of its vigilance. The
French hsemovigilance concept has been put into practice;
its network is now operational. Haemovigilance is the
ultimate quality indicator of blood transfusion and should
be part and parcel of the transfusion service. Its
organisation has to be adapted to the specific context of
each country. Likewise, haemovigilance will evolve with
the progress of health procurement organisations and a
better understanding of the benefits of quality assessments
by all those involved in transfusion therapy.
Aknowledgements
.................................................................................................................................................
The French Haemovigilance System
z
This presentation reflects the work of the entire hlemovigilance
network. The authors acknowledge F. Lepont and V Massari of
Unit6 Inserm U444 H6pital Saint-Antoine, J-P Aullen, M-P Vo Mai
and F. Sari of the hamovigilance cell at the AFS and the group of
the regional hlemovigilance co-ordinators for their contributions.
Vox Sang 1998; 74 (Suppl2) : 441-445
445