Journal of Thrombosis and Haemostasis, 7: 930–937 DOI: 10.1111/j.1538-7836.2009.03357.

IN FOCUS

Intensive peri-operative use of factor VIII and the

Arg593 fi Cys mutation are risk factors for inhibitor
development in mild/moderate hemophilia A
C.L. ECKHARDT,*1 L.A. MENKE,*1 C.H. VAN OMMEN,* J.H. VAN DER LEE,  R.B. GESKUS,à
P . W . K A M P H U I S E N , § M . P E T E R S * and K . F I J N V A N D R A A T *
*Department of Pediatric Hematology;  Department of Pediatric Clinical Epidemiology, Emma ChildrenÕs Hospital, Academic Medical Center,
Amsterdam; àDepartment of Clinical Epidemiology, Biostatistics and Bioinformatics; and §Department of Vascular Medicine, Academic Medical
Center, Amsterdam, the Netherlands.

To cite this article: Eckhardt CL, Menke LA, van Ommen CH, van der Lee JH, Geskus RB, Kamphuisen PW, Peters M, Fijnvandraat K. Intensive
peri-operative use of factor VIII and the Arg593 fi Cys mutation are risk factors for inhibitor development in mild/moderate hemophilia A.
J Thromb Haemost 2009; 7: 930–7.

See also Aledort LM. Mild and moderate factor VIII deficiency: inhibitor risks. This issue, pp 928–9.

inhibitor patients FVIII was administered by continuous

Summary. Background: A severe and challenging complica-
tion in the treatment of hemophilia A is the development of
inhibiting antibodies (inhibitors) directed towards factor VIII
(FVIII). Inhibitors aggravate bleeding complications, disabil-
ities and costs. The etiology of inhibitor development is
incompletely understood. Objectives: In a large cohort study in
patients with mild/moderate hemophilia A we evaluated the
role of genotype and intensive FVIII exposure in inhibitor
development. Patients/methods: Longitudinal clinical data
from 138 mild/moderate hemophilia A patients were retro-
spectively collected from 1 January 1980 to 1 January 2008 and
analyzed by multivariate analysis using Poisson regression.
Results: Genotyping demonstrated the Arg593Cys missense
mutation in 52 (38%) patients; the remaining 86 patients had
26 other missense mutations. Sixty-three (46%) patients
received intensive FVIII concentrate administration, 41 of
them for surgery. Ten patients (7%) developed inhibitors, eight
of them carrying the Arg593Cys mutation. Compared with the
other patients, those with the Arg593Cys mutation had a 10-
fold increased risk of developing inhibitors (RR 10; 95% CI,
0.9–119).The other two inhibitor patients had the newly
detected mutations Pro1761Gln and Glu2228Asp. In both
these patients and in five patients with genotype Arg593Cys,
inhibitors developed after intensive peri-operative use of FVIII
concentrate (RR 186; 95% CI, 25–1403). In five of the 10
Correspondence: Karin Fijnvandraat, Department of Pediatric
Hematology, (room G8-205), Academic Medical Center, P.O. Box
22660, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands.
Tel.: +31 20 566 2727; fax: +31 20 6917735.
E-mail: C.J.Fijnvandraat@amc.uva.nl
1
Both authors contributed equally to this study.
Received 9 January 2009, accepted 18 March 2009
infusion during surgery (RR 13; 95% CI, 1.9–86).
Conclusion: The Arg593Cys genotype and intensive peri-
operative use of FVIII, especially when administered by
continuous infusion, are associated with an increased risk for
inhibitor development in mild/moderate hemophilia A.
Keywords: antibody, genetics, hemophilia A, intensive FVIII
exposure, mild/moderate, surgery.
Introduction
A severe and challenging complication in the treatment of
hemophilia is the development of factor VIII (FVIII) inhibiting
antibodies (inhibitors). Inhibitors compromise the manage-
ment of bleeding symptoms, thereby aggravating complica-
tions, disabilities and costs [1,2]. Inhibitors arise more
frequently in severely affected patients than in those who are
mildly or moderately affected. The cumulative incidence of
inhibitors is estimated to be 23% in the former group and 8%
in the latter [3]. The lower incidence of inhibitor development in
mild and moderate hemophilia may be explained by the
circulating FVIII protein in these patients. Mild and moderate
hemophilia A is caused by missense mutations in the FVIII
gene and patients have a residual level of circulating endog-
enous FVIII activity [4]. This endogenous FVIII presumably
induces tolerance towards infused FVIII. In contrast, severely
affected hemophilia A patients have deletions or inversions of
the FVIII gene and no measurable FVIII in their plasma.
However, when inhibitors arise in patients with mild or
moderate hemophilia, the effects may be severe. These
inhibitors may not only neutralize the effect of FVIII replace-
ment therapy, but may also cross-react with endogenous
FVIII, causing severe spontaneous bleeding [5–7]. From a case
series of 26 patients with mild or moderate hemophilia who

 2009 International Society on Thrombosis and Haemostasis


developed inhibitors, two patients died from uncontrollable
hemorrhage [8]. As more than 50% of all hemophilia patients
have a mild or moderate form [9], the clinical impact of the
problem is substantial. A national study comprising a 22-year
observation period demonstrated that almost one-third of the
newly diagnosed inhibitors occurred in mild or moderate
hemophiliacs [10].
Most research on the etiology of inhibitor development has
been performed in patients with severe hemophilia A. Previous
studies in severe hemophilia A demonstrated that both genetic
and environmental factors play a role in the etiology of
inhibitor development [11].
Genetic factors associated with inhibitor development are
located within the FVIII gene (i.e. specific missense mutations
in regions encoding the A2 or C2 domain of the FVIII protein)
or outside the FVIII gene (e.g. African or Asian ethnic origin
and genetic polymorphisms in the IL-10 or TNFalfa gene)
[6,12–16]. Environmental factors that have been reported as
potential risk factors for inhibitor development are: early
intensive treatment, mode of FVIII administration (bolus
injections or continuous infusion), type of administered FVIII
concentrate, and change of FVIII products [6,17–23]. It has
also been suggested that infection, trauma and surgery may
render the immune system more susceptible to inhibitor
formation [19,20,24,25].
So far, reports on the possible association of certain risk
factors and inhibitor development in mild and moderate forms
of hemophilia only consist of case reports and case series, while
larger, comparative studies are absent. In order to evaluate the
potential role of genotype and intensive use of FVIII in the
development of inhibitors in mild or moderate hemophilia A,
we performed a longitudinal observational study in a well-
defined group of consecutive patients.
Patients and methods
Patient selection and genotyping
All mild [FVIII clotting activity (FVIII:C) 6–40 International
Units per milliliter (IU mL)1)] and moderate (FVIII:C 2–
5 IU mL)1) hemophilia A patients, who visited the hemophilia
treatment center of the Academic Medical Center on at least
three occasions during the studyÕs observation period were
eligible for this observational cohort study. The observation
period started on 1 January 1980, when intermediate purity
FVIII concentrates were introduced, and lasted until 1 January
2008. Patients who used no FVIII concentrate during the study
period were excluded. Patients were followed until inhibitor
development, death, or the end of the observation period.
Mutation analysis was performed, aided by pedigree analysis
for deceased patients for whom no DNA was available.
Data collection
Clinical data were collected from hospital databases and
patient files by two investigators (CLE, LAM), and discussed
 2009 International Society on Thrombosis and Haemostasis
FVIII inhibitors in mild/moderate hemophilia A 931
with two experienced hemophilia consultants (KF, MP)
whenever clarification was necessary. The number of exposure
days (ED) and the cumulative amount of FVIII concentrate
(CIU) administered during the study period were recorded. A
period of intensive use of FVIII concentrate was defined as the
cumulative use of at least 250 International Units per kilogram
bodyweight (IU kg)1) within five consecutive days, or at least
30 IU kg)1 day)1 during more than five consecutive days. The
reason for intensive use of FVIII was classified as being surgery
or a bleeding episode. Data collected on other potential risk
factors for inhibitor development included: ethnic origin, type
of FVIII product, product change and mode of FVIII
administration (bolus injections or continuous infusion).
Inhibitor development
FVIII:C was assessed by a one-stage clotting assay. Inhibitors
were identified by the Bethesda assay as described by Kasper
et al. [26]. From 1996 onward, the Nijmegen modification of the
Bethesda assay was used [27]. Patients were tested for inhibitors
when FVIII response declined after a period of intensive use of
FVIII, when bleeding tendency increased, or once a year if
FVIII concentrate had been used in the past 12 months. A titer
of at least 1 Bethesda Unit per milliliter (BU mL)1) was defined
as a low inhibitor titer, a titer of at least 5 BU mL)1 was defined
as a high inhibitor titer. If the result of the Bethesda assay was
between 0 and 1 BU mL)1, the patient was classified as an
inhibitor patient if the patient presented with spontaneous
bleeding symptoms, or if the FVIII ratio (FVIII:C during
inhibitor/FVIII:C before inhibitor) was 0.5 or less.
Statistics
In the description of patient characteristics continuous data are
presented as medians and interquartile ranges (IQR).
In order to evaluate the role of genotype, product change,
intensive use of FVIII, continuous infusion and surgical
procedures in the development of inhibitors, a Poisson
regression model was fitted. Inhibitor risk of individuals from
the same family was assumed to be correlated by including an
unobserved ÔfrailtyÕ or random effect for the event that was
assumed to have a normal distribution on the linear scale of the
predictors [28]. Potential risk factors with values changing over
time (intensive FVIII administration, continuous infusion and
product change) were entered as time-dependent covariates.
The effect of these covariates was assumed to last for 3 months
after the exposure. Results were presented as relative risks
(RR). Poisson regression analyses are performed in the R
statistical program [29].
Results
Patient characteristics, genotype and clinical risk factors
The study population consisted of 128 mild and 10 moderate
hemophilia A patients and comprised 6525 exposure days
932 C.L. Eckhardt et al
during an observation period of 1536 patient years. The median
period of observation was 10 years (IQR, 5–17). All but two
patients were of Caucasian origin. Genotyping demonstrated
the Arg593Cys missense mutation in 52 patients (38%); the
ancestors of the 17 families that these patients belonged to all
originated from the close vicinity of Amsterdam. In the
remaining 86 patients (62%), 26 other missense mutations were
found in 47 independent families. The characteristics of all
patients are depicted in Table 1.
During the study period, 63 patients (46%) received at least
one period of intensive FVIII concentrate administration. In 41
patients a surgical operation was the reason for the first
intensive FVIII administration. Twenty-two patients received
their first intensive FVIII concentrate administration for a
bleeding episode (Fig. 1). After this first intensive exposure,
another 61 periods of intensive FVIII concentrate administra-
tion were observed, of which 34 were for surgery and 27 for a
bleeding episode.
Inhibitor development
Inhibitor tests were performed in 131 patients (95%) with a
median of six inhibitor tests per patient (IQR, 3–8). The median
period between inhibitor testing was 13 months (IQR, 11–19).
In the other seven patients no inhibitor test was performed
during the study period. These patients had a low number of
exposure days to FVIII concentrates (median 2; IQR, 2–5) and
did not have an immediate indication for testing, such as
intensive exposure or increased bleeding tendency.
Ten patients (7%) developed inhibitors (Table 2). Four were
low titer and five were high titer inhibitors; the peak titer was
unknown in one patient. In six patients severe spontaneous
bleeding occurred, leading to death in one of them. This patient
died from uncontrollable hemorrhage and sepsis after immu-
nosuppressive treatment. Three other inhibitor patients died
from causes unrelated to their inhibitor.
All inhibitor patients were of Caucasian origin and the
median age at inhibitor development was 37 years (IQR, 18–
Table 1 Characteristics of the 138 mild and moderate hemophilia patients
at the end of the observation period
n (%) or median (IQR)
Total length of observation, years 1536
Total no. exposure days 6525
Age, years 41 (18–61)
Baseline FVIII:C, IU mL)1 14 (8–20)
Total no. of patients with the 52 (38%)
Arg593Cys mutation
Cumulative amount of 57 800 (8800–81 900)
FVIII administered, IU
No. exposure days 10 (6–29)
Type of FVIII product
Plasma derived 47 (34%)
Recombinant 91 (66%)
No. of product changes 1 (0–2)
No. inhibitor tests during the study period 6 (3–8)
No. months between inhibitor tests 13 (10–19)
64). Inhibitors arose after a median of 27 exposure days (IQR,
12–34), during which a median of 107 200 IU of FVIII
concentrate had been administered (IQR, 39 200–187 300). Six
of the 10 inhibitor patients had been treated with recombinant
FVIII concentrates prior to inhibitor development and the
other four had been treated with plasma-derived FVIII
concentrates.
The Arg593Cys mutation was present in eight inhibitor
patients; in the other two the newly detected missense
mutations Pro1761Gln and Glu2228Asp were found.
The incidence rate of inhibitor development was 13/1000
patient years in the Arg593Cys group, compared with 2/1000
patient years in the other mutation group (crude RR 6.5; 95%
CI, 3.5–18.5) (Table 3). The number of exposure days at the
end of the study was comparable for the Arg593group and the
other mutation group, median 10 ED (IQR 6–27) and median
10 ED (IQR 5–31), respectively, limiting the chance that the
results are confounded by the number of exposure days,
In seven patients (including five from the Arg593Cys
mutation group), inhibitors arose after intensive peri-operative
use of FVIII. Surgical procedures were: percutaneous coronary
intervention for coronary heart disease, total thyroidectomy for
thyroid carcinoma, total mesorectal excision for rectal cancer,
bilateral orchidectomy for testicular torsion, inguinal hernia
repair, tonsillectomy, and nose correction. Desmopressin
response had been tested in five of these seven patients and
FVIII: C response infusion varied from 38 to 93 IU dL)1.
All post-surgical inhibitors arose after the first surgical
procedure. Consequently, only the first surgical procedure per
patient was taken into account in the multivariate analysis.
The median number of days between the start of the first
peri-operative use of FVIII concentrate administration and the
detection of inhibitors was 47 (IQR, 26–70). Intensive peri-
operative use of FVIII concentrate was identified as a risk
factor for inhibitor development (RR, 186; 95% CI, 25–1403)
(Table 4). Intensive treatment with FVIII concentrates was not
only given for surgery but also for bleeding; this was the case in
22 patients. None of the twenty-two patients that received
intensive FVIII concentrate for bleeding episodes developed an
inhibitor (Fig. 1).
When the risk for inhibitor development in the Arg593Cys
mutation group was adjusted for intensive peri-operative use of
FVIII, continuous infusion, product change and family struc-
ture, the RR was 10 (95% CI, 0.9–119) (Table 4).
There was no significant difference between the surgical
patients and the patient group treated intensively for bleedings
with respect to the number of exposure days preceding
intensive treatment (median 2 ED, IQR 0–8 vs. median 2
ED, IQR 0–11 ED, respectively) or the number of exposure
days during intensive treatment (median 8 ED, IQR 7–10 vs.
median 7, IQR 5–11 ED, respectively) or the total dose of
FVIII concentrate (median 26 500 CIU, IQR 18 000–41 500
vs. median 21 500CIU, IQR 11 000–33 100) during the
intensive exposure. The seven patients who developed an
inhibitor after intensive peri-operative use of FVIII concen-
trates for surgery had a preceding FVIII exposure status of 12
 2009 International Society on Thrombosis and Haemostasis
 FVIII inhibitors in mild/moderate hemophilia A 933

A. 138 patients

B. Arg593Cys Other
mutations
(52) (86)

No intensive No intensive
C. Surgery Bleeding Surgery Bleeding
treatment treatment
(16) (8) (25) (14)
(28) (47)

D. CI No CI No CI No CI CI No CI No CI No CI
(6) (10) (8) (28) (4) (21) (14) (47)

E. Inhibitors Inhibitors Inhibitors Inhibitors Inhibitors Inhibitors Inhibitors Inhibitors

(4) (1) (0) (3) (1) (1) (0) (0)

Fig. 1. Flow chart of patients. Flow chart of all included patients (n = 138), classified as (B) those with and without the Arg593Cys mutation, and further
classified as (C) those who underwent intensive FVIII exposure for surgery or for bleeding and those who did not. Finally, patients were classified as
(D) those who received intensive FVIII exposure by continuous infusion (n = 10) or bolus injections. The number of inhibitors for each sub-classification
is shown in row E. Arg593Cys, Arg593Cys mutation in FVIII gene; Surgery, intensive FVIII exposure for surgery; Bleeding, intensive FVIII exposure
for bleeding without surgery; No intensive treatment, no intensive FVIII exposure during observation period; CI, continuous infusion of FVIII
concentrate; Inhibitors, inhibitor development.

Table 2 Characteristics of the patients who developed an inhibitor

FVIII:C
FVIII:C before during
Patient Preceding Type of Inhibitor peak inhibitor inhibitor FVIII:C Spontaneous
number Age Mutation event FVIII used  titer (BU mL)1) (IU dL)1) (IU dL)1) ratioà bleeding§ Outcome–

1 63 Arg593Cys Surgery MP >400 22 <2 <0.5 Y O

2 51 Arg593Cys Surgery RP 1.2 19 8 <0.5 N U
3 60 Arg593Cys Surgery RP 0.2 27 5 <0.5 N U
4 18* Arg593Cys Surgery IP 22 20 <2 <0.5 Y S
5 68 Arg593Cys Surgery RP 94 12 <2 <0.5 Y D
6 10 Arg593Cys 2 days of FVIII RP 1.7 15 5 <0.5 Y S
exposure for a
joint bleeding
7 23 Arg593Cys Infection RP 0.5** 17 4 <0.5 Y S
antibiotics
8 18* Arg593Cys None IP 9 20 3 <0.5 Y O
9 67 Pro1761Gln Surgery RP 9 6 <2 <0.5 N O
10 16 Glu2228Asp Surgery MP 1.6 17 18 1 N U

*Patients are HLA-identical twin brothers.

 
Type FVIII concentrate used at time of inhibitor development. MP, monoclonal purified plasma derived concentrate; IP, intermediate purity
plasma derived concentrate; RP, recombinant product.
à
FVIII:C ratio, FVIII:C during inhibitor/FVIII:C before inhibitor.
§
Y, yes; N, no.
–
Outcome, clinical manifestation of inhibitor; D, patient died, death was related to inhibitor; O, patient died, death was not related to inhibitor; S,
severe spontaneous bleeding; U, unchanged bleeding pattern.
**Highest measured titer, however, inhibitor peak titer unknown due to a long-term stay abroad.

 2009 International Society on Thrombosis and Haemostasis

934 C.L. Eckhardt et al

Table 3 Crude (i.e. not corrected for family structure) incidence rates of putative risk factors and inhibitor development by univariate analysis
Exposed to risk factor Unexposed to risk factor

Putative risk factor No. inhibitors* PYAR  No. inhibitors* PYAR

à
Arg593Cys mutation 8 621 2 915à
First intensive FVIII treatment for surgery 7 10 3 1526
Continuous infusion of FVIII concentrates 5 1.9 5 1534
FVIII product change 1 61 9 1475
*
Total number of patients who developed an inhibitor during the 28-year observation period.
 
PYAR, person-years at risk.
à
Person-years at risk, person-years observed (PYO), because mutation is a time-independent covariate.

Table 4 Multivariate survival analysis of putative risk factors of inhibitor
development, adjusted for family structure
No.
exposures Adjusted
to risk No. RR
factor inhibitors (95% CI)  P
Arg593Cys mutation 52 8 10 (0.9–119) 0.06
First intensive FVIII 41 7 186 (25–1403) £0.0001*
treatment for surgeryà§
Continuous infusion of 10 5 13 (1.9–86) 0.008*
FVIII concentratesà§
FVIII product changeà 220 1 1.1 (0.1–11.7) 0.9
*The significance level was set at P < 0.05.
Each variable was adjusted for all the others and for family structure,
to adjust for confounding and correlated data.
à
In a period of 3 months prior to inhibitor development.
§
For calculating the relative risk of intensive use of FVIII and con-
tinuous infusion of FVIII concentrates only the first surgical proce-
dures were taken into account, as all post-surgical inhibitors arose
after the first surgical procedure.
ED (IQR, 8–24), which is higher than the median of two ED in
the total group of patients receiving intensive FVIII treatment
for surgery. Hence, it seems unlikely that the observed
association between intensive peri-operative use of FVIII
concentrates and inhibitor development was confounded by a
low number of exposure days prior to surgery.
In five (50%) of the inhibitor patients FVIII was adminis-
trated by continuous infusion (Fig. 1). Continuous infusion
was used in 10 patients in this study, exclusively in surgical
patients and not for treatment of bleeding episodes. The high
incidence (5/10) of inhibitors after continuous infusion sug-
gested an association between continuous infusion of FVIII
concentrates and inhibitor development. The adjusted relative
risk of continuous infusion for inhibitor development was 13
(95% CI, 1.9–86) (Table 4).
In the remaining three inhibitor patients, all Arg593Cys
genotype, inhibitor development was not preceded by a period
of intensive FVIII exposure. The first patient had been treated
with FVIII concentrate (50 IU kg)1 day)1) for a joint bleed in
his ankle for 2 days prior to inhibitor detection. The second
patient was suffering from recurrent infections combined with
the use of antibiotics shortly before inhibitor development. For
the third patient, no remarkable event had taken place in the
6 months prior to inhibitor detection. He had received FVIII
replacement therapy on about five occasions before he
developed an inhibitor.
Half of the patients (n = 70) switched FVIII product for a
total of 220 times. In seven patients this product change took
place at the moment of surgery. Only one patient developed an
inhibitor within 3 months after changing FVIII product; he
switched from a recombinant product to another recombinant
product on the day he was operated on and developed an
inhibitor 64 days later. (Table 2; patient no 3). In the whole
study population there was no significant relation between
FVIII product change and inhibitor development (RR, 1.1;
95% CI, 0.1–11.7) (Tables 3 and 4).
Discussion
This first cohort study on the etiology of inhibitors in patients
with mild or moderate hemophilia strongly suggests that
intensive peri-operative use of FVIII concentrate, especially
when delivered by continuous infusion, and to a lesser extent
the Arg593Cys mutation, are risk factors for the development
of inhibitors in this group. No inhibitors occurred after
intensive use of FVIII concentrate for bleeding. The number of
inhibitor patients was too small to investigate the possible
interaction of the Arg593Cys mutation and intensive peri-
operative FVIII exposure. Two new missense mutations were
observed in the patients who developed inhibitors after
intensive peri-operative use of FVIII: Pro1761Gln and
Glu2228Asp.
Previously, case reports have been published about inhibitor
development in patients with Arg593Cys missense mutation
[6,30–35]. These reports concern three patients from the current
study (Table 2, patients number 1, 4 and 8), one patient from
Canada [33] and two patients from Sweden [34]. No previous
study has addressed the causative role of the mutation in a
cohort study or case control study. The mutation leads to
replacement of the amino acid arginine by cysteine at position
593 in the A2 domain of the FVIII protein. Because cysteine is
involved in the formation of disulphide bridges, its presence
may result in a conformational change of the FVIII protein.
Therefore, it may alter the immunogenic characteristics of the
endogenous FVIII. Upon the administration of wild-type
FVIII, this might predispose for inhibitor development [31,35].

 2009 International Society on Thrombosis and Haemostasis


The pedigrees of the Arg593Cys patients in this study
suggested a founder effect. Because other genetic factors than
the mutation in the FVIII gene can contribute to the risk of
inhibitor development, it can not be excluded that shared
immunological genes other than the Arg593Cys mutation
might play a role in the observed risk for inhibitor development
in these patients [14,15]. Therefore we adjusted for family
structure in the multivariate analysis. However, except for two
HLA identical twin brothers, the patients in whom inhibitors
developed were not closer than fourth to eleventh degree family
members. Because all but two patients were of Caucasian
origin, we could not investigate whether ethnicity is a risk
factor for inhibitor development in mild or moderate hemo-
philia.
Both intensive peri-operative use of FVIII and continuous
infusion have been mentioned as possible risk factors for
inhibitor development in patients with mild or moderate
hemophilia [17,18]. The risk of intensive peri-operative use
of FVIII was demonstrated in severe hemophilia A patients
in the CANAL cohort study [23], but it has never been
studied in mild or moderate hemophilia A before. As no
inhibitors developed after intensive treatment with FVIII for
bleeding, tissue damage that results from surgery may play
an important pathophysiological role by triggering the
immune system in such a way that it is more prone to
develop inhibitors [24,36,37]. When interpreting the in-
creased risk found for continuous infusion on inhibitor
development the low number of patients that received FVIII
by continuous infusion should be taken into account.
Nevertheless, our results strongly suggest an attributive risk
of continuous infusion during surgery for inhibitor devel-
opment. The antibody response against exogenous FVIII
in these patients may be explained by the theoretical
release of danger signals elicited by subcutaneous leakage
of FVIII concentrate during continuous infusion. Other
factors possibly contributing to this increased risk are the
modification of FVIII protein during storage in infusion
pumps, or concomitant thrombophlebitis at the infusion site
[17–19].
However, as two patients who developed inhibitors post-
surgically had exclusively received bolus injections, the risk of
post-surgical inhibitors could not be attributed solely to the use
of continuous infusion.
In our cohort, all post-surgical inhibitors occurred after the
first surgical procedure, suggesting that especially this first
exposure to surgery combined with intensive use of FVIII
selects those patients who are more susceptible to develop
inhibitors. As all post-surgical inhibitors were detected within
3 months after surgery, this period was used as a time window
for this covariate in the Poisson regression model. The
extremely high relative risk found for intensive peri-operative
use of FVIII causing inhibitor development is partially due to
this narrow time window during which the variable is assumed
to be effective; nevertheless the association of intensive peri-
operative use of FVIII with inhibitor development seems
strong.
 2009 International Society on Thrombosis and Haemostasis
FVIII inhibitors in mild/moderate hemophilia A 935
Although inhibitor development after changing FVIII
product has been suggested in mild hemophilia A [21], until
now there has been no evidence that a change of product is
associated with an increased inhibitor risk [38]. In our cohort
only one patient developed an inhibitor after product change.
As this product change was combined with intensive peri-
operative FVIII exposure on the same day, the risk attributable
to product change could not be discerned.
Several studies have demonstrated that age at first FVIII
exposure is inversely associated with inhibitor risk in severe
hemophilia A [23,39–41]. Because of the advanced age of our
cohort, many patients had been treated with cryoprecipitate or
other blood products before FVIII concentrates became
available. Because reliable data could not be retraced, we did
not analyze FVIII concentrate that was administered before
1980 or prescribed outside the study center, and the influence of
age at first FVIII exposure on inhibitor development could not
be analyzed. The reported cumulative number of exposure days
and CIU may underestimate the total lifetime exposure,
although we expect that this occurred randomly and that the
observed relations are not confounded.
To summarize, the intensive use of FVIII concentrate in
surgical procedures and the Arg593Cys mutation are both
associated with a high risk of inhibitor development in mild or
moderate hemophilia A. In order to prevent the severe
complication of developing inhibitors, we propose that this
risk should be taken into account when treating patients with
mild and moderate hemophilia A intensively with FVIII
concentrates for surgical procedures, particularly in those
carrying high risk mutations.
Hence, surgery should only be performed for clear indica-
tions.
Our results imply that limiting intensive FVIII exposure or
alternative therapeutic options, such as the additional use of
DDAVP in mild hemophilia A, should be considered whenever
possible [42–44].
Addendum
Contribution: C.L. Eckhardt and L.A. Menke performed
research, collected and analyzed data and wrote the paper;
C.H. van Ommen performed the mutation analysis and
critically reviewed the paper; J.H. van der Lee supervised data
analysis and writing of the paper; R.B. Geskus analyzed data
and critically reviewed the paper; P.W. Kamphuisen supervised
writing of the paper and critically reviewed the paper; M. Peters
supervised data collection and writing of the paper;
K. Fijnvandraat designed research, supervised data collection
and data analysis, and drafted and wrote the paper.
Disclosure of Conflict of Interests
K. Fijnvandraat is a member of the European Hemophilia
Treatment and Standardization Board sponsored by Baxter.
The other authors state that they have no conflicting financial
interest.
936 C.L. Eckhardt et al
References
1 Franchini M, Salvagno GL, Lippi G. Inhibitors in mild/moderate
haemophilia A: an update. Thromb Haemost 2006; 96: 113–18.
2 Colowick AB, Bohn RL, Avorn J, Ewenstein BM. Immune tolerance
induction in hemophilia patients with inhibitors: costly can be cheaper.
Blood 2000; 96: 1698–702.
3 Wight J, Paisley S. The epidemiology of inhibitors in haemophilia A: a
systematic review. Haemophilia 2003; 9: 418–35.
4 Jacquemin M, De Maeyer M, DÕOiron R, LavendÕHomme R,
Peerlinck K, Saint-Remy JM. Molecular mechanisms of mild
and moderate hemophilia A. J Thromb Haemost 2003; 1: 456–63.
5 Jacquemin M, Vantomme V, Buhot C, Lavend’Homme R, Burny W,
Demotte N, Chaux P, Peerlinck K, Vermylen J, Maillere B, van der
Bruggen P, Saint-Remy JM. CD4+ T-cell clones specific for wild-type
factor VIII: a molecular mechanism responsible for a higher incidence
of inhibitor formation in mild/moderate hemophilia A. Blood 2003;
101: 1351–8.
6 Hay CR. Factor VIII inhibitors in mild and moderate-severity hae-
mophilia A. Haemophilia 1998; 4: 558–63.
7 Franchini M, Girelli D, Olivieri O, Castaman G, Lippi G, Poli G,
Salvagno GL, Tagariello G, Giuffrida A, de Gironcoli M, Morfini M,
Berntorp E, Gandini G. Tyr2105Cys mutation in exon 22 of FVIII
gene is a risk factor for the development of inhibitors in patients with
mild/moderate haemophilia A. Haemophilia 2006; 12: 448–51.
8 Hay CR, Ludlam CA, Colvin BT, Hill FG, Preston FE, Wasseem N,
Bagnall R, Peake IR, Berntorp E, Mauser Bunschoten EP, Fi-
jnvandraat K, Kasper CK, White G, Santagostino E. Factor VIII
inhibitors in mild and moderate-severity haemophilia A. UK Hae-
mophilia Centre Directors Organisation. Thromb Haemost 1998; 79:
762–6.
9 Plug I, van der Bom JG, Peters M, Mauser-Bunschoten EP, de Goede-
Bolder A, Heijnen L, Smit C, Zwart-van Rijkom JE, Willemse J,
Rosendaal FR. Thirty years of hemophilia treatment in the Nether-
lands, 1972–2001. Blood 2004; 104: 3494–500.
10 Darby SC, Keeling DM, Spooner RJ, Wan Kan S, Giangrande PL,
Collins PW, Hill FG, Hay CR, UK Haemophilia Centre DoctorsÕ
Organisation. The incidence of factor VIII and factor IX inhibitors in
the hemophilia population of the UK and their effect on subsequent
mortality, 1977–99. J Thromb Haemost 2004; 2: 1047–54.
11 Oldenburg J, Schroder J, Brackmann HH, Muller-Reible C, Schwaab
R, Tuddenham E. Environmental and genetic factors influencing
inhibitor development. Semin Hematol 2004; 41: 82–8.
12 Aledort LM. Hemophilia. N Engl J Med 2001; 345: 1066–7.
13 Astermark J, Berntorp E, White GC, Kroner BL. The Malmo Inter-
national Brother Study (MIBS): further support for genetic predispo-
sition to inhibitor development in hemophilia patients. Haemophilia
2001; 7: 267–72.
14 Astermark J, Oldenburg J, Pavlova A, Berntorp E, Lefvert AK.
Polymorphisms in the IL10 but not in the IL1beta and IL4 genes are
associated with inhibitor development in patients with hemophilia A.
Blood 2006; 107: 3167–72.
15 Astermark J, Oldenburg J, Carlson J, Pavlova A, Kavakli K, Berntorp
E, Lefvert AK. Polymorphisms in the TNFA gene and the risk of
inhibitor development in patients with hemophilia A. Blood 2006; 108:
3739–45.
16 Scandella D, Gilbert GE, Shima M, Nakai H, Eagleson C, Felch M,
Prescott R, Rajalakshmi KJ, Hoyer LW, Saenko E. Some factor VIII
inhibitor antibodies recognize a common epitope corresponding to C2
domain amino acids 2248 through 2312, which overlap a phospho-
lipid-binding site. Blood 1995; 86: 1811–19.
17 Batorova A, Martinowitz U. Continuous infusion of coagulation
factors. Haemophilia 2002; 8: 170–7.
18 von Auer C, Oldenburg J, von Depka M, Escuriola-Ettinghausen C,
Kurnik K, Lenk H, Scharrer I. Inhibitor development in patients with
hemophilia A after continuous infusion of FVIII concentrates. Ann N
Y Acad Sci 2005; 1051: 498–505.
19 Sharathkumar A, Kirby M, Freedman M, Abdelhaleem M, Chitayat
D, Teshima IE, Dror Y. Intensive exposure to factor VIII is a risk
factor for inhibitor development in mild hemophilia A. J Thromb
Haemost 2003; 1: 1228–36.
20 White B, Cotter M, Byrne M, OÕShea E, Smith OP. High responding
factor VIII inhibitors in mild haemophilia–is there a link with recent
changes in clinical practice? Haemophilia 2000; 6: 113–15.
21 Baglin T, Beacham E. Is a change of factor VIII product a risk factor
for the development of a factor VIII inhibitor? Thromb Haemost 1998;
80: 1036–7.
22 Yee TT, Lee CA. Is a change of factor VIII product a risk factor for
the development of a factor VIII inhibitor? Thromb Haemost 1999; 81:
852.
23 Gouw SC, van der Bom JG, Marijke van den Berg H. Treatment-
related risk factors of inhibitor development in previously untreated
patients with hemophilia A: the CANAL cohort study. Blood 2007;
109: 4648–54.
24 Koestenberger M, Leschnik B, Muntean W. More on: mild hemophilia
A and inhibitor development. J Thromb Haemost 2004; 2: 676.
25 Gouw SC, van den Berg HM, le Cessie S, van der Bom JG. Treatment
characteristics and the risk of inhibitor development: a multicenter
cohort study among previously untreated patients with severe hemo-
philia A. J Thromb Haemost 2007; 5: 1383–90.
26 Kasper CK, Aledort L, Aronson D, Counts R, Edson JR, van Eys J,
Fratantoni J, Green D, Hampton J, Hilgartner M, Levine P, Lazerson
J, McMillan C, Penner J, Shapiro S, Shulman NR. Proceedings: a
more uniform measurement of factor VIII inhibitors. Thromb Diath
Haemorrh 1975; 34: 612.
27 Verbruggen B, Novakova I, Wessels H, Boezeman J, van den Berg M,
Mauser-Bunschoten E. The Nijmegen modification of the Bethesda
assay for factor VIII:C inhibitors: improved specificity and reliability.
Thromb Haemost 1995; 73: 247–51.
28 Aalen OO. Heterogenity in survival analysis. Stat Med 1988; 7: 1121–
37.
29 R Development Core Team (2008). R: a Language and Environment for
Statistical Computing. R Foundation for Statistical Computing,
Vienna, Austria. ISBN 3-900051-07-0. http://www.R-project.org;
accessed 9 January 2009.
30 Fijnvandraat K, Turenhout EA, van den Brink EN, ten Cate JW, van
Mourik JA, Peters M, Voorberg J. The missense mutation
Arg593 fi Cys is related to antibody formation in a patient with mild
hemophilia A. Blood 1997; 89: 4371–7.
31 van den Brink EN, Timmermans SM, Turenhout EA, Bank CM,
Fijnvandraat K, Voorberg J, Peters M. Longitudinal analysis of factor
VIII inhibitors in a previously untreated mild haemophilia A
patient with an Arg593 fi Cys substitution. Thromb Haemost 1999;
81: 723–6.
32 Bril WS, Turenhout EA, Kaijen PH, van den Brink EN, Koopman
MM, Peters M, Voorberg J. Analysis of factor VIII inhibitors in a
haemophilia A patient with an Arg593 fi Cys mutation using phage
display. Br J Haematol 2002; 119: 393–6.
33 Thompson AR, Murphy ME, Liu M, Saenko EL, Healey JF, Lollar P,
Scandella D. Loss of tolerance to exogenous and endogenous factor
VIII in a mild hemophilia A patient with an Arg593 to Cys mutation.
Blood 1997; 90: 1902–10.
34 Knobe KE, Villoutreix BO, Tengborn LI, Petrini P, Ljung RC. Factor
VIII inhibitors in two families with mild haemophilia A: structural
analysis of the mutations. Haemostasis 2000; 30: 268–79.
35 Bril WS, MacLean PE, Kaijen PH, van den Brink EN, Lardy NM,
Fijnvandraat K, Peters M, Voorberg J. HLA class II genotype and
factor VIII inhibitors in mild haemophilia A patients with an Arg593
to Cys mutation. Haemophilia 2004; 10: 509–14.
36 Grzelak I, Olszewski WL, Zaleska M, Durlik M, Lagiewska B, Mus-
zynski M, Rowinski W. Blood cytokine levels rise even after minor
surgical trauma. J Clin Immunol 1996; 16: 159–64.
37 Robbins D, Kulkarni R, Gera R, Scott-Emuakpor AB, Bosma K,
Penner JA. Successful treatment of high titre inhibitors in mild
 2009 International Society on Thrombosis and Haemostasis

hemophilia A with avoidance of factor VIII and immunosuppressive
therapy. Am J Hematol 2001; 68: 184–8.
38 Gouw SC, van der Bom JG, Auerswald G, Ettinghausen CE, Tedgard
U, van den Berg HM. Recombinant versus plasma-derived factor VIII
products and the development of inhibitors in previously untreated
patients with severe hemophilia A: the CANAL cohort study. Blood
2007; 11: 4693–7.
39 van der Bom JG, Mauser-Bunschoten EP, Fischer K, van den Berg
HM. Age at first treatment and immune tolerance to factor VIII in
severe hemophilia. Thromb Haemost 2003; 9(Suppl 1): 75–81.
40 Santagostino E, Mancuso ME, Rocino A, Mancuso G, Mazzucconi
MG, Tagliaferri A, Messina M, Mannucci PM. Environmental risk
 2009 International Society on Thrombosis and Haemostasis
FVIII inhibitors in mild/moderate hemophilia A 937
factors for inhibitor development in children with haemophilia A: a
case-control study. Br J Haematol 2005; 130: 422–7.
41 Lorenzo JI, Lopez A, Altisent C, Aznar JA. Incidence of factor VIII
inhibitors in severe haemophilia: the importance of patient age. Br J
Haematol 2001; 112: 600–3.
42 Aledort LM. Can we prevent inhibitors in mild and moderate factor
VIII-deficient patients? J Thromb Haemost 2006; 4: 914.
43 DÕOiron R, Pipe SW, Jacquemin M. Mild/moderate haemophilia A:
new insights into molecular mechanisms and inhibitor development.
Haemophilia 2008; 14(Suppl 3): 138–46.
44 Mannucci PM. Desmopressin (DDAVP) in the treatment of bleeding
disorders: the first twenty years. Haemophilia 2000; 6(Suppl 1): 60–7.