The development of factor VIII (FVIII) inhibiting antibodies (inhibitors) is the most severe complication of treatment with clotting factor concentrates for hemophilia A. Inhibitors compromise the ability to manage hemorrhage in affected patients, resulting in a considerable increase in complications, disability and costs [2,3]. About one in four severe, and one in 15 non-severe hemophilia A patients develop inhibitors during their treatment . The reason why these patients develop inhibitors is not completely known, and this hampers the development of an effective strategy to reduce the risk.
The etiology of inhibitor development is a complex process in which multiple genetic and environmental factors interact dynamically . Patients who develop inhibitors are likely to have high-risk genotypes. Inhibitor development is triggered by certain environmental factors during their treatment, such as intensive treatment with clotting factor concentrates, inflammation and infection [6,7]. Inflammation may provoke antibody formation by the concurrent presence of cytokine release arising from injured tissues, so called ‘danger signals’ .
Although the association between intensive treatment and the formation of inhibiting antibodies towards FVIII in hemophilia A has been demonstrated, it is presently unclear through which mechanism this severe complication is elicited . Moreover, the influence of reason for treatment (surgery vs. bleeding) is indistinct. Not only may tissue damage and injury elicit immunological danger signals during surgery, several other factors such as anesthetic drugs and inflammation may contribute to antibody formation. In addition, administration of FVIII concentrates by continuous infusion has been suggested to contribute to a higher incidence of inhibitors perioperatively. Several conditions related to continuous infusion, such as subcutaneous leakage of FVIII concentrate, FVIII protein modification during storage in infusion pumps, or concomitant thrombophlebitis at the infusion site, may lead to danger signal release and thereby activation of the immune system .
Perioperative FVIII replacement regimens are targeted to prevent bleeding and do not take the potential inhibitor risk into account. Over the last few decades this has resulted in a tendency to aim for higher FVIII levels, leading to the use of higher doses of factor concentrates in surgical procedures . If inhibitor risk of intensive treatment for surgical interventions could be predicted in individual patients, clinicians would be able to adapt their management of these patients to reduce the risk.
Although numerous reports on intensive treatment with FVIII concentrates in combination with surgery as a risk factor for inhibitor development in hemophilia A have been published, no systematic review of the literature is available. By combining data of individual studies, the inclusion of a high number of patients will lead to the critical mass that is needed to quantify more accurately the effect of this risk factor.
The purpose of this systematic review was to address the contribution of surgery to inhibitor formation in patients with hemophilia A receiving intensive treatment with FVIII concentrates.
Therefore we formulated three research questions, based on contrasting the effect of high and low doses of FVIII (a) and contrasting the effect of the presence of immunological danger signals with their absence (b), as illustrated in Fig. 1.
- 1 Is there an association between intensive treatment with FVIII concentrates for surgery and inhibitor development as compared with prophylaxis? (a) High vs. low load of FVIII antigen. (b) Presence vs. absence of danger signals.
- 2 Is there an association between intensive treatment with FVIII concentrates for surgery and inhibitor development as compared with intensive treatment for a bleeding episode? (a) High vs. intermediate load of FVIII antigen. (b) Higher vs. lower presence of danger signals.
- 3 Is continuous administration of FVIII concentrates during intensive treatment for surgery associated with a higher risk of inhibitor development as compared with administration by bolus infusions? (a) Both high load of FVIII antigen. (b) Higher vs. lower presence of danger signals.
We conducted a systematic review and meta-analysis according to the Meta-analysis of Observational Studies in Epidemiology (referred to as ‘MOOSE’) Group checklist . This MOOSE consensus statement was developed to provide uniform guidance for the conduct of meta-analysis of observational studies, in order to increase their quality.
We comprehensively searched the literature of CENTRAL, MEDLINE and EMBASE (January 1966–December 2010) using terms for hemophilia (e.g. Haemophil* a, Hemophil* a, Haemophiliac* or Hemophiliac*), factor VIII (e.g. factor VIII*, FVIII*, or factor 8*), inhibitor (e.g. inhibit* or antibody*) and epidemiology (Epidemiolog*, epidemiologic factors). The full list of terms is available as supplemental online data. Abstracts of empirical studies that appeared potentially relevant were identified independently by two of the authors (CE and KF) by checking the records retrieved by the electronic search. For those studies that were potentially relevant, full papers were obtained and studies were selected for inclusion. Additional articles were identified by reviewing reference lists. Searches were not restricted to language. The manufacturers of FVIII concentrates (Bayer Healthcare, Mijdrecht, the Netherlands; Baxter, Utrecht, the Netherlands; CSL Behring, Breda, the Netherlands; Wyeth Pharmaceuticals, Hoofddorp, the Netherlands) were contacted for additional data and information regarding unpublished data of FVIII treatment for surgery.
Randomized controlled trials, cohort studies and case control studies were included if the participants were hemophilia A patients (irrespective of age, severity or treatment history), if treatment for surgery, bleeding and/or prophylaxis was reported, and if the outcome was inhibitor development. Studies were included when either all participants were previously untreated patients at the beginning of the observation, or if previously treated participants had never been tested positive for inhibitors and the number of previous exposure days (EDs) to FVIII concentrates were reported. Surgery was defined as an invasive medical procedure involving an incision of skin or mucosa with instruments, or closure of a previously sustained wound. Any surgical procedure was considered, irrespective of urgency, type of procedure, body system involved, degree of invasiveness, and type of instrumentation. Other procedures were considered as surgery (e.g. angioplasty or endoscopy) if they involved ‘common’ surgical procedure or settings, such as use of a sterile environment, anesthesia, antiseptic conditions, typical surgical instruments, and suturing or stapling. Intensive treatment with FVIII concentrates was defined as the use of FVIII concentrates during at least 3 days. We defined inhibitor development as at least one positive inhibitor assay (≥ 0.6 BU mL−1) during the study observation period.
Studies were excluded if there was no contrast of study populations within the study (i.e. determinant series [all patients underwent surgery] or case-series [all patients developed an inhibitor] or case-reports). Also studies including only patients with acquired hemophilia A or female carriers of hemophilia A were excluded.
Two authors (CE and MN) independently extracted data from full articles of the included studies. If needed, the authors of the primary studies were contacted for clarification of data and additional information. The assessments were carried out without any masking. A standardized data extraction form was used to retrieve data of interest, including study characteristics, participant characteristics, treatment prior to and during the study period, continuous/bolus infusion, measure of inhibitor development, follow-up, and potential confounders.
Assessment of methodological quality
Five authors (CE, JB, PK, MP and KF) assessed the methodological quality of the studies using a modified quality assessment scale, each study being independently reviewed by two individuals. The modified quality assessment scale was based on earlier described checklists for observational studies according to evidence-based medicine criteria [13,14].
The assessment of methodological quality concerned the following items: comparability of groups, selection bias, follow-up, outcome, blinding and confounders. Confounders considered in each study were: age, preceding number of EDs, previous intensive treatment and/or surgery, product type and product switch, ethnicity, positive family history of inhibitors, FVIII genotype and severity of disease. The methodological quality assessment criteria for observational studies are described in Table 1. For the methodological quality assessment of case–control studies, the criteria were slightly adapted with regard to the selection of cases and controls. Cases and controls had to be selected based on comparable patient characteristics (i.e. age, number of prior EDs to FVIII concentrates and FVIII mutation). Discrepancies between reviewers regarding choice of articles meriting inclusion, data extraction and quality assessment were resolved by discussion.
|Complete||If type and extent of surgery was clear, together with description of exposure to FVIII concentrates before (cumulative exposure days before surgery), during and following surgery for all patients who underwent surgery during the study period. |
And if mode of administration of factor VIII concentrates (bolus or continuous infusion) during surgery is clear.
And if choice and type of control group is clear (treatment for bleeding or prophylactic treatment with FVIII concentrates), together with description of exposure to FVIII concentrates.
|Cohort study: if the group consisted of consecutive or obviously representative series of hemophilia A patients (> 90% of the original cohort of hemophilia A patients). |
Or if it was an appropriate random sample with respect to the previous exposure to FVIII concentrates was taken.
Case–control study: for the risk of bias assessment of case–control studies, the criteria are adapted with regard to the selection of cases and controls.
Cases and controls have to be selected from the same source population (preferably stratified on number of previous exposure days). Source population is defined as follows: ‘if the controls would have become a case, they would have been selected in the current study as a case-patient.’
|If there was follow-up of at least 3 months after intensive treatment with factor VIII concentrates for surgery or bleeding or after prophylactic treatment in all patients. |
And if inhibitor tests (a minimum of two tests) were performed until the end of the follow-up period.
If the outcome was assessed at the end date of the study for more than 90% of the study group of interest.
And if there was a complete follow-up (> 3 months) of all subjects accounted for.
|If the outcome definition was objective and precise (i.e. if all patients underwent inhibitor tests during the study period). |
And if there was a statement of no history of inhibitors for all patients.
And if inhibitor levels have been measured within 3 months after intensive treatment for surgery/bleeding or prophylactic treatment.
|If other important risk factors (age, number of preceding exposure days, previous intensive treatment and surgery, product type and product switch and genetic risk factors: ethnicity, positive family history of inhibitors, FVIII genotype and severity of disease) have been identified and association with inhibitor development has been analyzed for these factors in univariate analysis. |
And if important prognostic factors (age, number of exposure days and previous intensive treatments/surgeries) and follow-up were taken adequately into account.
And if relative risk, odds ratio, attributable risk, linear or logistic regression model, or mean difference, was calculated for more than 90% of the population.
|Adequate||If surgery status is described for all patients.And exposure before and during the perioperative period is described.||Cohort study: if eligible HA patients over a defined period of time, in a defined catchment area, or in a defined hospital or clinic, group of hospitals or health maintenance organization were included. |
Case–control study: if all patients were drawn from the same community or source independent of factor exposure history.
|If time of follow-up is explicitly stated in the method section on the study design. |
And if time of follow-up is adequate considering the specific aim of the study.
And if the outcome was assessed at the end date of the study for 80% of the study group of interest.
Or if description was provided of those lost and if subjects lost to follow-up were unlikely to introduce bias.
|If description was provided of those who were not tested and there was no clinical likelihood of inhibitor presence (i.e. no spontaneous bleeding symptoms and no decreased response to FVIII concentrates).||If cumulative number of preceding exposure days, severity of disease and FVIII mutation are mentioned for all patients. |
And if a multivariate analysis has been conducted, but not with all above-mentioned prognostic factors and/or follow-up.
We used Review Manager (Review Manager 5.0.15, 2008, The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark). for data analyses. Because of the use of observational data and the risk of bias, only studies that specifically aimed to study the risk of surgery for inhibitor development were included in the meta-analysis. Incidence of inhibitor development in patients that received intensive treatment with FVIII concentrates for surgery was compared with the incidence of inhibitor development in others who were treated with FVIII concentrates for prophylaxis or bleeding. Raw dichotomous data were calculated as odds ratios (ORs) or Relative Risks (RRs) with 95% confidence intervals (CIs) and were pooled using the Mantel Haenzel method. A random effects model was employed because heterogeneity was assumed to exist, based on the observational character of the included studies.
Our initial search yielded 2494 literature citations. After exclusion of 382 duplicates, 2112 unique references remained (Fig. 2). Of these, 1961 were excluded after scanning titles or reading the abstracts. Of the 151 unique articles retrieved, 128 were excluded after full article review: in 60 studies there were no data on surgery or no surgery was performed during the observation period, 24 studies were case reports or case series in which all patients underwent surgery (23) or all had an inhibitor (1), in five studies inhibitors pre-existed before study-entry, in 21 studies information on inhibitor outcome was not reported, four were reviews, eight were comments and/or contained no original data, two were experimental reports, and four studies were published in a non-English language (Spanish, Japanese, Chinese and Hungarian) and the full article could not be obtained.
Twenty-three studies were considered for inclusion in the systematic review [15–37]. An attempt was made to contact all corresponding authors for lacking data; three authors and one pharmaceutical company provided additional data before 1 January 2011 [25,29,34]. After thorough examination of the 23 articles, 16 were excluded from the systematic review because they did not aim to study the association between surgery and inhibitor development in hemophilia A. Consequently, chronological data about prophylaxis, bleeding and surgery for each patient were lacking in these 16 studies, hampering the analysis of these determinants in cases and controls. The remaining seven studies [15,16,20,25,26,34,37], comprising 957 hemophilia A patients, were included in the systematic review. Methodological quality was assessed for all seven studies and was adequate to complete for most of them. The results of the quality assessment are given in Table 2.
|Eckhardt et al. ||+||++||++||++||+|
|Gouw et al. ||+||++||++||++||++|
|Gouw et al. ||−||+||+||++||++|
|Kempton et al. ||−||+||−||−||+|
|Maclean et al. ||+||+||++||++||−|
|Santagostino et al. ||+||++||++||++||+|
|Sharathkumar et al. ||+||++||++||++||−|
Table 3 presents the characteristics of the seven included studies. Four of the seven studies were cohort studies [15,20,25,26] and three were case–control studies [16,34,37]. Four studies (two cohort, two case–control) included only patients with a severe form of hemophilia A (FVIII:C ≤ 1%) [16,25,26,34], and the other three (two cohort, one case–control) included only non-severe hemophilia A patients (FVIII:C 2–40%) [15,16,20,37].
|Eckhardt et al. ||Cohort||Median 10 EDs (IQR 6–29)||128 mild, |
|41 (18–61) years‡||NS||47 (34%) plasma, |
91 (66%) rec‡
|138 PUP||138/10||NS/NS||75/7||Factor VIII product change, continuous infusion, family structure, Arg593Cys mutation, intensive treatment for bleeding.|
|Gouw et al. ||Cohort||50 EDs||366 severe||11 (6–15) months§, |
31 (21–46) months‡
|24 (15%)||NS||366 PUP||366/87||17/11||84/NS||Baseline FVIII activity level, ethnicity, FVIII gene mutation type, age at first exposure, duration between EDs, dose, prophylaxis, peak treatment moment at first treatment, reason for first treatment and product type.¶|
|Gouw et al. ||Cohort||50 EDs||272 severe||9 (6–13) months§, |
38 (27–52) months‡
|26 (11%)||272 (100%) rec||236 PUP||236/67||4/2||44/NS||Baseline FVIII activity level, ethnicity, family history of inhibitors, age at first exposure and prophylaxis.**|
|Kempton et al. ||Case–control||48 pat < 50 EDs, |
50 pat > 50 EDs
|48 Mild |
|Controls: 31 years |
Cases: 31 yrs‡
|11 (11%)||17 (17%) plasma, |
66 (67%) rec‡
|Maclean et al. ||Case–control||> 50 EDs||156 Severe||Cases: 284 days |
Controls; 289 days§
|14 (9%)||45 (29%) plasma, |
111 (71%) rec
|Santagostino et al. ||Case–control||> 50 EDs||108 severe||Cases: 11 (range 2 days–64 months) |
Controls: 13 (1 day –67 months)§
|13 (12%)||NS||108 PUP||48/60††||2/7††||11/15††||NS|
|Sharathkumar et al. ||Cohort||Mean 16 ED (range 1–46)||54 Mild||5.4 years (range 5 days to 16.3 years)§||NS||4 (14%) plasma |
Type of surgical procedure varied greatly among the inhibitor patients; the most performed procedures were dental extractions, orthopedic procedures and catheter implantations. The studies did not specify type of surgical procedures in non-inhibitor patients. Only one study mentioned cumulative FVIII dose and number of EDs perioperatively .
A total of 342 patients developed an inhibitor during the study observation period. The cut-off for a positive inhibitor assay ranged from > 0.5 BU mL−1  to ≥ 0.6 BU mL−1 [25,26,37] and ≥ 1 BU mL−1 [16,20] and was not stated in one study . The frequency of inhibitor testing was defined as ≥ 4/year [16,25,34] or ≥ 4/50 EDs  or 1/year or after intensive treatment , and was not stated in two studies [15,37].
Inhibitor risk of surgery at first exposure
Data on treatment at first exposure were available in the four studies on patients with severe hemophilia A, two cohort studies and two case–control studies [16,25,26,34]. In the multicenter cohort of 366 severe hemophilia A patients reported by Gouw and colleagues , 65% (11/17) of the patients that were first treated for surgical procedures developed an inhibitor compared with 22% (65/286) of the patients who were treated for bleeding and 23% (8/37) of the patients who received prophylaxis at first treatment. For the other 26 patients the reason for first exposure to FVIII and inhibitor development related to this first exposure was not stated in the article. Adjusted RR of surgical procedure at first treatment for inhibitor development was 2.6 (CI, 1.3–5.1).
The other study by Gouw and colleagues  describes a cohort of 236 severe previously untreated patients derived from databases of four FVIII product registration studies. Data on treatment at first exposure were available for 234 (99%) patients. Half of the patients (2/4, 50%) that received their first FVIII concentrates for surgery developed an inhibitor, compared with 25% (45/180) of the patients that were treated for bleeding and 44% (7/16) who received prophylaxis at first treatment (data obtained from authors). For the other 34 patients the reason for first treatment was recovery (eight patients) or other, not further specified (26 patients); 13 of them developed an inhibitor (38%). Based on these data, we did not find a statistically significant association between surgery as the reason for first exposure and inhibitor development (OR, 2.7; CI, 0.4–20.2).
The case–control study by Santagostino and colleagues  included 60 inhibitor patients and 48 controls with moderate or severe hemophilia A. Surgery was the reason for starting FVIII treatment in seven cases (12%), all with high-responding inhibitors, and two (4%) controls. There was no statistically significant association between surgery compared with bleeding at first exposure and inhibitor development (crude OR, 3.0; CI, 0.6–15.4).
The other case–control study by Maclean and colleagues  included 78 inhibitor cases and 78 age-matched controls. Four of the cases (5%) and two of the controls (3%) had surgery as the reason for their first exposure. Compared with minor bleeds, there was no significant association between surgery and inhibitor development (crude OR, 2.2; CI, 0.4–12). In both case–control studies the chronological order of events was not analyzed nor was surgery included in multivariate analysis [16,34].
Based on these four studies [16,25,26,34], crude pooled OR for inhibitor development after surgery compared with non-surgery (i.e. bleeding or prophylaxis) at first exposure was 4.1 (CI, 2.0–8.4) (Fig. 3).This association remained when surgery was compared with bleeding alone (crude OR, 4.1; CI, 2.0–8.5). In only two studies [25,26] was prophylactic treatment given as the reason for first exposure. Surgery was associated with a 2-fold increased risk of inhibitor development compared with prophylaxis (crude pooled RR, 2.1; CI, 0.8–5.2).
Inhibitor risk of surgery in previously treated patients
The four studies in patients with severe hemophilia A also reported data on surgery and inhibitor development in (minimally) previously treated patients [16,25,26,34]. In the cohort study by Gouw and colleagues  25% (80/366) of patients underwent 84 major surgical procedures, defined as surgery for which replacement therapy was given on at least three consecutive days. Major surgical procedures on any ED (including surgery at first exposure) were associated with a slightly increased risk of developing inhibitors (adjusted RR, 1.4; CI, 0.8–2.5) as compared with the period before surgery. Relative risk for inhibitor development after portacath implantations (adjusted RR, 1.4; CI, 0.7–2.7) was comparable to the relative risk of other surgical procedures (adjusted RR, 1.3; CI, 0.5–3.0).
In the other study by Gouw and colleagues  major surgical procedures during the first 50 EDs were associated with a higher risk (RR, 2.4; CI, 1.2–4.8) of developing clinically relevant inhibitors as compared with the period prior to surgery.
In both case–control studies no association between surgery on any ED and inhibitor development was present in univariate analysis [16,34]. Because of the potential influence of time-varying confounding determinants (e.g. number of cumulative EDs, number of preceding surgeries or treatment intensity) on the association of surgery on any ED and inhibitor development, we decided that pooling of individual studies was not justified.
Inhibitor risk and treatment intensity
In three studies in severe patients, treatment intensity was analyzed as a potential risk factor irrespective of reason for treatment [16,25,26]. In the two cohort studies two different classifications of treatment intensity were used according to the moment in time that intensive treatment took place: at first exposure or at any exposure [25,26]. The intensity of treatment at first exposure was classified into three categories: < 3 consecutive EDs, 3–4 consecutive EDs and ≥ 5 EDs (defined as intensive treatment). For treatment intensity on any ED two categories of treatment intensity were defined: peak treatment moments defined as treatment on ≥ 3 EDs, and major peak treatment moments defined as treatment on ≥ 5 EDs.
In the first cohort study intensive treatment at first exposure was associated with a more than 3-fold increased risk (adjusted RR, 3.3; CI, 2.1–5.3) . This association remained for peak treatment moments and major peak treatment moments on any ED but was less pronounced (peak treatment: adjusted RR, 1.5; CI, 0.9–2.5; major peak treatment: adjusted RR, 1.6; CI, 1.0–2.6).
In the other cohort study the association between intensive treatment at first exposure and inhibitors was confirmed (adjusted RR, 2.7; CI, 1.2–5.8) . Again, in this study the associations of peak treatment and major peak treatment with inhibitor development were less pronounced (respectively, adjusted RR, 1.6; CI, 1.0–2.6; and adjusted RR, 1.6; CI, 0.9–2.8).
Maclean and colleagues  classified treatment intensity into four categories: < 3 EDs; and three cumulative categories, ≥ 3 EDs (including the two following categories), ≥ 5 EDs (including the following category) and ≥ 10 EDs. Compared with non-intensive treatment (< 3 EDs), the association of treatment intensity of at least 3, 5 or 10 EDs and inhibitor development was increasingly pronounced with unadjusted ORs of, respectively, 1.7 (CI, 0.8–3.5), 2.4 (CI, 0.9–6.3) and 5.0 (CI,1.03–24).
Pooled results of these three studies showed that intensive treatment of at least three consecutive EDs at first exposure is associated with inhibitor development (crude OR, 2.1; CI, 1.2–3.7) compared with < 3 EDs [16,25,26]. This association was more pronounced when intensive treatment of ≥ 5 EDs was compared with < 3 EDs (crude OR, 4.1; CI, 2.6–6.5) (Fig. 4). None of the three studies reported reasons for treatment intensity or adjusted for reasons for intensive treatment (i.e. bleeding or surgery), when analysing the association between treatment intensity and inhibitor development.
Inhibitor risk of surgery in patients with mild and moderate hemophilia A
We identified three studies on surgery and inhibitor development in non-severe patients; two cohort studies [15,20] and one case–control study . In a cohort of 29 mild hemophilia patients described by Sharatkumar and colleagues, 14% (4/29) of the patients developed inhibitors . All inhibitors occurred after intensive treatment with FVIII concentrate, defined as more than six consecutive EDs. In two patients inhibitor development was preceded by intensive treatment for surgery. The total number of surgical procedures in the other 25 non-inhibitor patients was not described; therefore association between surgery or intensive treatment and inhibitor development could not be calculated.
In the other cohort study 10 of the 138 patients (7%) developed an inhibitor . Of these 10 patients, seven (70%) were treated intensively with FVIII concentrate for surgery prior to inhibitor development. Intensive exposure was defined as at least five consecutive EDs with a cumulative FVIII use of at least 250 units per kilogram bodyweight or 30 IU kg−1day−1. If the reason for the first intensive exposure was a surgical procedure this was associated with a 186-fold increased risk of inhibitor development (CI, 25–1403) in the 3 months after the surgical procedure as compared with the period in which no intensive treatment with FVIII was given. After the first intensive exposure period, 36 patients received one or more subsequent intensive treatments for surgery or bleeding. None of these patients developed an inhibitor.
In the case–control study 36 inhibitor cases and 62 controls were included . Half of the inhibitor cases (18/36, 50%) and 18% (11/62) of the controls received intensive FVIII treatment during the prior year, defined as six or more consecutive days of FVIII replacement. Intensive treatment in the prior year was strongly associated with inhibitor development (OR, 4.6; CI, 2.8–11.7). Surgery was the indication for intensive treatment in 78% (14/18) of the cases, compared with 36% (4/11) of controls. The association between surgery and inhibitor development was not further analyzed.
In these three studies in mild and moderate hemophilia A patients information on the chronological order of events was not presented. [15,20,37] Because the contribution of these unknown factors to inhibitor development cannot be quantified, no pooled analysis of the association between surgery and inhibitor development in non-severe patients could be made.
Continuous infusion vs. bolus injections
Mode of administration during surgery was reported in the three studies including non-severe hemophilia A patients [15,20,37]. In the cohort study by Sharatkumar and colleagues seven of the 16 patients (44%) that underwent intensive treatment with FVIII concentrates received continuous infusions . All four inhibitors occurred after continuous infusion (4/7 = 57% of all patients receiving continuous infusion). In the other cohort study continuous infusion was used in 10 of the 41 patients (25%) during intensive treatment for their first major surgery . Five of these patients (50% of all patients receiving continuous infusion) developed an inhibitor afterwards. After adjustment for surgery, intensive treatment, product change, FVIII genotype and family structure, the RR of continuous infusion for inhibitor development was 13 (CI, 1.9–86) in the 3 months following surgery covered by continuous infusion as compared with the period in which no intensive treatment with FVIII was given.
In the case–control study in non-severe patients by Kempton and colleagues  continuous infusion of FVIII was used during the period of intensive FVIII treatment in seven of the 18 cases (39%) and in two of the 11 controls (18%). No significant association was found between continuous infusion and inhibitor development as compared with bolus injections (P = 0.41, OR and CI not mentioned). Because of heterogeneity of the studies and missing data we were not able to pool data from the studies to perform a meta-analysis.
Our study shows that surgery in combination with FVIII treatment is strongly associated with inhibitor development in hemophilia A as compared with intensive treatment for bleeding or prophylaxis alone. This association is especially pronounced if surgery is the reason for first treatment and seems present both in severe and non-severe hemophilia A. Information on continuous infusion was scarce.
We comprehensively searched the literature for relevant articles following systematic review methods. The use of robust methodological procedures to estimate the risk of bias of potentially relevant articles strengthens our conclusions. The quality of retrieved studies was not always optimal, as information on prior exposure status of the patients, type of surgical procedure and chronological association between treatment and inhibitor development was frequently lacking. To reduce potential risk of bias, we restricted the meta-analysis to studies specifically focussing on risk factors for inhibitor development. As these studies consisted of representative series of hemophilia A patients covering a population heterogeneous in origin, age and severity of disease, we assume that this inclusion criterion has not introduced severe selection bias.
Although our meta-analysis was limited to the use of crude risk estimates, the possibility of confounding seems to be limited, as adjustment for time-varying determinants in the two studies by Gouw and colleagues  did not lead to major changes in risk assessments (major surgical procedures at any exposure, crude RR 2.4 vs. adjusted RR 2.7; surgical procedures, crude RR 3.7 vs. adjusted RR 2.6 ).
The association between intensive treatment > 5 EDs and inhibitor development was previously acknowledged in the CANAL study . However, treatment for surgical procedures is in general more intensive than FVIII treatment for bleeding episodes. It is therefore extremely difficult to study surgery and intensive treatment independent of each other. None of the included studies provided detailed information on the reason for intensive treatment. Another limitation was that actual levels of FVIII were not reported in the studies that we reviewed and could not be included in the meta-analysis.
Previous studies showed that the risk of developing inhibitors is strongly associated with the number of previous exposures to FVIII . In severe patients half of the inhibitors occur before the 15th ED and the other half occur with a sharply decreasing incidence rate relatively early afterwards. After 50 EDs the risk of developing new inhibitors is < 1%. This review showed that especially intensive treatment for surgery at first treatment is associated with inhibitor development (OR 4.1). Surgery at any exposure within the first 50 EDs was also associated with inhibitor development but less pronounced (RR 1.4–2.4). Thus, a higher number of EDs prior to surgery may reduce the risk of inhibitor development.
Prophylaxis is already widely used as standard treatment regimen in severe hemophilia, based on favourable joint outcome resulting from decreased bleeding frequency . However, evidence is lacking on the optimal timing of initiation of primary prophylaxis . From an immunological point of view , the opportunity should be taken to obtain tolerance in a ‘danger-free’ environment and therefore prophylaxis should be initiated before exposure related to surgery or a major bleed. Treatment was initially started as prophylaxis in a small proportion of patients (7–11%) in the two studies that were included in the meta-analysis [25,26]. This may be due to the fact that most treatment regimens in severe patients start prophylaxis after the first bleeding episode (joint bleed or cerebral bleeding). Based on these two studies the pooled OR for inhibitor development following treatment for surgery in comparison to prophylaxis was remarkably less pronounced (OR 2.1) then when surgery was compared with bleeding (OR 4.1). As we used non-randomized data for unadjusted meta-analysis, this may be explained by the clinical selection of patients for initiation of prophylaxis before any intensive treatment period. Further research is needed to evaluate whether the protective effect of prophylaxis can be confirmed.
In non-severe hemophilia A patients prophylactic treatment is rarely given because these patients have a less severe bleeding pattern . The presence of low amounts of circulating FVIII protects these patients from spontaneous bleeding and therefore FVIII concentrate is mostly administered for surgery or a major bleed. Minor bleeds or interventions are preferably covered by desmopressin (DDAVP) in mild patients with a good response . FVIII is predominantly administered when immunological danger signals are present, when surgery is performed or major bleeds occur. This may well explain the relatively high incidence of inhibitors at a more advanced age in non-severe patients after intensive treatment periods (Table 3).
In the article by Kempton and colleagues , 42% of the non-severe hemophilia A patients developed inhibitors after more than 50 EDs, suggesting that non-severe patients are at risk of inhibitor development for a longer time period than severe patients. Absence of information on previous intensive exposures, including previous surgical interventions, precluded estimation of the effect of surgery on inhibitor development in non-severe patients. Certain missense mutations (e.g. Arg593Cys and Arg2150Cys) in mild hemophilia patients are associated with a higher risk of inhibitor formation . Well-designed further studies are needed in non-severe hemophilia A patients to estimate the risk of surgical interventions for inhibitor development, taking the influence of other genetic and environmental risk factors into account. This will ultimately enable implementation of individual treatment strategies in non-severe patients who are at high risk of inhibitor development.
Considering the consistent finding of increased inhibitor incidence associated with surgical intervention across a number of heterogeneous patient groups, we believe that caution must be used when treating patients intensively with FVIII concentrates for surgery, especially previously untreated patients. We recommend that intensive treatment with FVIII concentrates should be avoided early in treatment whenever possible. Starting low-dose prophylactic treatment at an early stage may provide an opportunity to induce tolerance in a ‘danger free’ environment .
Because of the increased life-expectancy of hemophilia patients, which will increase the incidence of age-related health problems such as cancer and arthrosis, a continued rise in the need for surgery in hemophilia A patients is to be expected [44,45]. Therefore, risk estimation in previously treated patients and mild/moderate hemophilia A patients, who might face their first intensive exposure to FVIII at an advanced age, should receive more attention. As intensity of treatment is associated with the risk of inhibitor development, ways to lower the amount or duration of FVIII replacement should be investigated. In mild hemophilia A patients, alternative or additional use of desmopressin should be considered to reduce intensive treatment.
Two studies in mild and moderate patients show a clear association between continuous infusion and inhibitor development. Storage conditions, phlebitis at the infusion site or other concomitant factors during continuous infusion may provoke the immune system to develop inhibitors. However, the number of patients that received continuous infusion in these studies was small [15,20]. Therefore, recommendations concerning the use of continuous infusion to cover surgical procedures or bleeding episodes should await further investigations.
This systematic review shows that surgery in combination with intensive FVIII treatment is strongly associated with inhibitor development in hemophilia A patients as compared with treatment for bleeding or prophylaxis. This review highlights the need for robust well-designed future studies on the association between surgery and inhibitor development in non-severe hemophilia A patients and previously treated patients, in which potential confounding factors are taken into account. Also the role of continuous infusion should be further elucidated. This knowledge will ultimately help us to understand the pathophysiology of surgery causing inhibitor development and may enable further research on preventive strategies, such as prophylaxis or immunosuppressive therapy.
C.L. Eckhardt designed and performed the research, collected and analyzed data and wrote the paper. J.G. van der Bom and K. Fijnvandraat designed research, supervised data collection and data analysis, performed quality assessment, and drafted and wrote the paper. M. van der Naald performed the research, collected and analyzed data and prepared the tables. M. Peters and P.W. Kamphuisen performed quality assessment and critically reviewed the paper.
Classic hemophilia or hemophilia A is a congenital bleeding diathesis in which the affected individual may present with spontaneous hemorrhage or persistent bleeding even after minor trauma. Knowledge about the disease process, multidisciplinary team approach, and timely management can lead to favorable outcome in these patients. We report management of a child with hemophilia A for suturing of lacerated upper lip mucosa following trauma. A review of literature with recommendations for perioperative management, especially in the setting of emergency surgery, is also provided.
Keywords: Anesthesia, emergency surgery, factor VIII, hemophilia, trauma
Classic hemophilia or hemophilia A is an X-linked recessive hereditary disorder characterized by defective or deficient clotting factor VIII (FVIII) which affects the male progeny. The incidence is 1/5000 male live births. The affected individual may present with spontaneous hemorrhage specifically into the joint cavities or uncontrollable bleeding even after minimal injury. In pediatric age group, the child's physical activity increases as the age advances which results in more exposure to trauma and bleeding. In toddlers, laceration of tongue and mouth is a common presentation due to biting of tongue or lip during a fall or trauma. Profuse bleeding from open wounds in children with hemophilia A results in significant blood loss because of delayed formation of soft and friable clot. Moreover, rebleeding or delayed bleeding during the physiologic lysis of clots can also occur after a period of apparent hemostasis. Hence, trivial trauma can sometimes lead to life-threatening hemorrhage in hemophilia patients. So, local as well as systemic approach is essential to manage such patients till the complete healing occurs. Exogenous administration of FVIII concentrate temporarily replaces the missing clotting factor. Hence, prompt transfusion of recombinant FVIII concentrate and early surgical intervention can transmogrify the outcome in these patients.
A 5-year-old, 17 kg boy was brought in emergency with excessive bleeding for 30 min following injury while playing, involving mucosa of the upper lip. At 1.5 years of age, he was incidentally diagnosed of having pink Tetralogy of Fallot for which he underwent surgical repair. The patient was also diagnosed as a case of hemophilia A since then. The patient's factor replacement history revealed moderate disease (2% FVIII activity) with a negative inhibitor titer assay within the last month and two previous episodes of acute bleeding following fall from height which was successfully treated with FVIII concentrate (Hemofil-M, Baxter). He had no previous history of spontaneous bleeding into joints; no other members of the family had any clinical disease.
On examination, his pulse rate was 130 beats/min, blood pressure was 84/50 mm of Hg. As the bleeding did not respond to pressure maneuvers (compression with damp gauge pieces), hematologist opinion was sought for and transfusion of recombinant FVIII concentrate was started in the emergency after securing a peripheral line with 22-gauge cannula. Since the child was not cooperative, suturing under monitored anesthesia care was planned. High-risk consent for surgery and anesthesia was obtained from the parents. Inside the operation theater, another 22-gauge peripheral line was secured and packed red blood cells transfusion was also started. Monitoring included pulse oximetry, electrocardiogram, and noninvasive blood pressure. He was preoxygenated with nasal prong at 4 L/min and neck was turned on one side and intermittent suction was done on dependent cheek with a red rubber catheter [Figure 1]. Injection ranitidine 1 mg/kg, injection ondansetron 0.15 mg/kg, and injection glycopyrrolate 8 µg/kg were administered intravenously. Infusion of loading dose of injection tranexamic acid 15 mg/kg was started before the induction of anesthesia. Anesthesia was induced with slow intravenous ketamine 2 mg/kg. Suturing of the bleeding mucosa with resorbable suture was performed and hemostasis achieved.
Simple (red) rubber catheter with blunt tip and side hole.
During surgery, meticulous suctioning and mopping were done to prevent aspiration of blood. Postoperatively, he was transferred to recovery room and nursed in the left lateral position with oxygen via face mask till he became fully conscious. Eight hours after surgery, the patient received another dose of FVIII concentrate and thereafter twice a day for 5 days. The postoperative period was uneventful. After 7 days, he was discharged from the hospital.
A child with hemophilia A presenting for emergency surgery poses a challenge to the surgeon as well as anesthesiologist. Bleeding tendency in the affected individual is inversely proportional to FVIII level in the body. Normal plasma levels of FVIII ranges from 0.5 to 1.5 IU/ml or 50–150% (1 IU/ml = 100% of FVIII in 1 ml of normal plasma), and the hemostatic level is more than 30–40%. FVIII activity in patients with mild, moderate, and severe hemophilia A are 5-<40%, 1–4%, and <1% of normal, respectively. Coagulation studies in patients with hemophilia A reveal high activated partial thromboplastin time (APTT) but normal platelet count, bleeding time, and prothrombin time. All patients with classic hemophilia, regardless of the severity of the disease, are at risk of excessive bleeding after trauma or during surgery.
Intracranial hemorrhage accounts for major cause of mortality in hemophilia patients. Besides that, morbidity and mortality in hemophilia patients are also increased due to bleeding from injury, gastrointestinal or genitourinary tract.
The indicated coagulation defect should be corrected as soon as possible and must not be delayed while results of diagnostic tests are awaited. The surgery in patients with hemophilia A should take place in a center with adequate laboratory support for perioperative assessment of FVIII level including inhibitor screening and optimal blood bank support to provide sufficient quantities of FVIII concentrates. As our patient needed emergency surgery, FVIII assay and inhibitor screening were not feasible preoperatively. Postoperative monitoring for bleeding was done with hemoglobin and APTT levels along with FVIII level and FVIII antibody assay.
Emergency bleeding incidents such as in case of trauma or emergency surgery require immediate intervention with a major dose of FVIII concentrate infusion to ensure hemostasis. Moreover, the consumption of coagulation factors is highly increased during surgery. Hence, the replacement of recombinant FVIII concentrate must occur before or in parallel with any intervention in a patient with hemophilia. The recommended plasma factor level and duration of administration for major as well as minor surgeries are shown in Table 1. Life-threatening hemorrhages require replacement therapy to achieve a plasma level equal to that of normal, i.e., 100 IU/dl or 100%. The dose of FVIII concentrate is calculated according to the World Federation of Hemophilia 2012 (WFH) guidelines: Patient's body weight (in kg) × desired factor level (IU/dl) × 0.5 or weight (in pound)/4.4 × desired factor level (IU/dl). In the absence of an inhibitor, plasma FVIII level increases approximately 2 IU/dl per infused IU/kg body weight. As FVIII assay was not available to us in emergency, we started slow IV infusion of 850 units (17 kg × 100 IU/dl × 0.5) of FVIII at 100 units per min considering the severe bleeding from mucosal trauma to achieve desired factor level of 100% as per the hematologist's advice. Since the half-life of FVIII is approximately 8–12 h, it must be administered twice daily.
WFH recommendation of desired factor VIII levels and duration of factor VIII administration for major and minor surgeries
Desmopressin, a synthetic vasopressin analog, increases the plasma levels of FVIII and Von Willebrand factor in patients with mild to moderate hemophilia A. A single IV infusion (0.3 μg/kg diluted in 50–100 mL of isotonic saline and infused over 30 min) increases the level of FVIII 3–6 times with peak response in 90 min. However, it is ineffective in patients with severe hemophilia A or patients with high FVIII antibody titers.
Tranexamic acid and ε-aminocaproic acid are useful as an adjunct therapy in hemophiliacs and are valuable in controlling bleeding from mucosal surface as it promotes clot stability by their antifibrinolytic activity. Tranexamic acid can be used as IV infusion, oral tablet, 4.8% solution as mouth wash, and application of solution or paste of crushed tablets directly to the site of bleeding.
Although vascular access does not cause excessive bleeding, it is better to place it with care and intramuscular medications and arterial puncture should be avoided. Smooth induction of anesthesia is preferable and succinylcholine is avoided to prevent muscle fasciculation, which may worsen muscle and joints hemorrhagic state. Postinduction manipulation or intubation of the airway can cause submucosal hemorrhages, which can become a life-threatening condition and nasal intubation is more traumatic and bleeding from the site can also lead to aspiration. Hence, tubes should be previously greased with lubricant to decrease friction with the mucosa and proper care is needed with the insertion of feeding tubes or temperature probes because tongue and airway muscles bleeding may rapidly lead to airway obstruction. We avoided tracheal intubation in our patient to prevent any insult to the airway mucosa. We placed the child supine with head turned on one side and continuous suctioning from dependent cheek was done with a nontraumatic red rubber catheter to prevent aspiration of secretions and blood. Pressure points were padded with cotton during positioning of the extremities to prevent intramuscular hematomas or hemarthrosis.
Hemodynamic conditions should be maintained as close as possible to normal range during anesthesia as hypertension and tachycardia can result in increased bleeding from surgical area. Tracheal extubation needs to be performed in deep plane of anesthetic without cough reflex with extreme vigilance about pharyngeal aspiration. Although guidelines do not recommend regional anesthesia in patients with hemophilia, literature review revealed that neuraxial blocks as well as peripheral nerve blocks have been performed safely in a patient with moderate to severe hemophilia with maintenance of FVIII level within a safe range throughout the perioperative period.
For closure of the mucosal wound, absorbable suture-like vicryl was used in our patients as it avoids need for the postoperative removal and the possibility of bleeding during suture removal.
Postoperatively, we used paracetamol for analgesia as nonsteroidal anti-inflammatory drug can predispose Hemophilia patients to gastrointestinal complications like ulcer, bleeding and perforation.
To conclude, we successfully managed a child with classic hemophilia with sublabial oral mucosal injury for emergency suturing. Management according to WFH guidelines, prompt surgery and less manipulation of airway possibly led to good outcome in our case.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
1. Kar A, Phadnis S, Dharmarajan S, Nakade J. Epidemiology & social costs of haemophilia in India. Indian J Med Res. 2014;140:19–31.[PMC free article][PubMed]
2. Rayen R, Hariharan VS, Elavazhagan N, Kamalendran N, Varadarajan R. Dental management of hemophiliac child under general anesthesia. J Indian Soc Pedod Prev Dent. 2011;29:74–9.[PubMed]
3. Kashyap R, Choudhry VP. Hemophilia. Indian Pediatr. 2000;37:45–53.[PubMed]
4. Scott JP, Flood VH. Hereditary clotting factor deficiencies (Bleeding Disorders) In: Kliegman RM, Stanton BF, Geme JW, Schor NF, Behrman RE, editors. Nelson Textbook of Pediatrics. 20th ed. Vol. 2. Philadelphia: Saunders, Elsevier; 2015. p. 2384.
5. Sahu S, Lata I, Singh S, Kumar M. Revisiting hemophilia management in acute medicine. J Emerg Trauma Shock. 2011;4:292–8.[PMC free article][PubMed]
6. Mallya P, Kaimar P, Jithesh R, Ranjan RK, Ambareesha M. Anaesthetic management of a patient with haemophilia. Indian J Anaesth. 2007;51:145.
7. Srivastava A, Brewer AK, Mauser-Bunschoten EP, Key NS, Kitchen S, Llinas A, et al. Guidelines for the management of hemophilia. Haemophilia. 2013;19:e1–47.[PubMed]
8. Gyanesh P, Dhiraaj S. Anesthetic management of a patient with hemophilia A with spontaneous acute subdural hematoma. J Anaesthesiol Clin Pharmacol. 2013;29:117–20.[PMC free article][PubMed]
9. Franchini M, Lippi G. The use of desmopressin in acquired haemophilia A: A systematic review. Blood Transfus. 2011;9:377–82.[PMC free article][PubMed]
10. Madan N, Rathnam A, Bajaj N. Treatment of an intraoral bleeding in hemophilic patient with a thermoplastic palatal stent – A novel approach. Int J Crit Illn Inj Sci. 2011;1:79–83.[PMC free article][PubMed]
11. Flores RP, Bagatini A, Santos AT, Gomes CR, Fernandes MS, Molon RP. Hemophilia and anesthesia. Rev Bras Anestesiol. 2004;54:865–71.[PubMed]
12. Zetlaoui PJ. Ultrasound guided axillary brachial plexus catheter in a patient with severe hemophilia. J Anesth Clin Res. 2013;4:321.
13. Jover-Cerveró A, Poveda Roda R, Bagán JV, Jiménez Soriano Y. Dental treatment of patients with coagulation factor alterations: An update. Med Oral Patol Oral Cir Bucal. 2007;12:E380–7.[PubMed]
14. Riley RR, Witkop M, Hellman E, Akins S. Assessment and management of pain in haemophilia patients. Haemophilia. 2011;17:839–45.[PubMed]