SEARCH

SEARCH BY CITATION

Keywords:

  • haemophilia A;
  • human factor VIII protein;
  • antibodies;
  • VWF ;
  • immunology

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

This systematic review was designed to summarize the reported valid quantitative evidence on the association between use of von Willebrand factor (VWF)-containing Factor VIII (FVIII) concentrates and successful immune tolerance induction (ITI) in patients with severe haemophilia A. The primary outcome was successful ITI; secondary outcomes were time to success, complications of the inhibitor or ITI and relapse of the inhibitor. A systematic literature search identified 26 randomized controlled trials, registries and cohort studies, evaluating a total of 1284 patients. For a pooled meta-analysis, 13 studies evaluating 382 patients were included. Due to incomplete data we were not able to assign pre-ITI risk categories to all patients for risk factor analysis. The meta-analysis did not demonstrate a difference in the proportion of patients with successful inhibitor eradication between those treated with VWF-containing products and those treated with FVIII concentrates devoid of VWF (relative risk [RR] 0·70 (95% confidence interval [CI] 0·52–0·89) and 0·84 (95% CI 0·75–0·93) respectively). Bleeding rate during ITI ranged from 0·00 to 0·85 bleeding episodes per year. The proportion of patients with a relapse of the inhibitor (range 0–20%) was mentioned in four studies that were included in the meta-analysis. The results of this systematic review do not support the idea of a positive effect of VWF-containing products in ITI.

The development of inhibiting antibodies (inhibitors) against Factor VIII (FVIII) is the most severe and challenging complication in the treatment of haemophilia with FVIII concentrates (Fijnvandraat et al, 2012). Inhibitors occur in 30% of the severe haemophilia A patients (Ehrenforth et al, 1992; Wight et al, 2003). Inhibitors may be eradicated by frequent administration of high doses of FVIII, a regimen referred to as immune tolerance induction (ITI). ITI is successful in about 65% of the patients (Mariani et al, 1994; Kreuz et al, 2003; Astermark et al, 2006; Berntorp et al, 2006; Mauser-Bunschoten et al, 2007; Dimichele, 2009). The optimal ITI regimen is the subject of intense discussions. A randomized trial compared low-dose (50 iu/kg three times/week) with high-dose (200 iu/kg/d) ITI in a good-risk cohort of severe haemophilia A patients with high-titre inhibitor (Hay & Dimichele, 2012). The success rate was similar between treatment arms; however the trial was terminated prematurely due to an increased number of bleeding events in the low-dose arm. Therefore, the study lacked the statistical power to demonstrate therapeutic equivalence below the 30% boundary of equivalence. Another highly debated issue is the type of FVIII concentrate to be used for ITI. In most cases, the same product that elicited the inhibitor is used. However, it has been proposed that von Willebrand factor (VWF)-containing FVIII concentrates may be more effective in achieving immune tolerance (Ettingshausen & Kreuz, 2005; Kaveri et al, 2011; Kurth et al, 2011). The concept that VWF-containing FVIII concentrates are more effective in inducing immune tolerance was initially based on positive reports from a single institution (Ettingshausen & Kreuz, 2005). Later, in vitro studies were published explaining that potential inhibitor epitopes are masked by VWF, thereby shielding the infused FVIII from rapid degradation and prolonging antigen presentation (Ettingshausen & Kreuz, 2005). Other in vitro studies have shown that inhibitors directed against the A2 domain and light-chain of FVIII are less active against VWF-bound FVIII in comparison to FVIII that does not contain VWF (Suzuki et al, 1996; Gensana et al, 2001). This may further contribute to improved success in inhibitor eradication for VWF-containing FVIII concentrates (Berntorp, 2001). Clinical studies addressing the role of VWF in ITI have yielded conflicting results. Some studies did not find an association between the type of concentrate and success of ITI, whereas others suggested that switching to VWF-containing concentrates can revert to a successful outcome after initial failure of ITI (Auerswald et al, 2003; Orsini et al, 2005; Gringeri et al, 2007; Greninger et al, 2008; Kurth et al, 2008; Valentino et al, 2009). This heterogeneity may be due to the small patient numbers in individual studies. A meta-analysis combining the data of the available studies may generate more robust estimates of the association of VWF-containing concentrates and successful ITI.

Several reviews have been published addressing this topic, however none of these reviews were designed as a systematic review, according to the definition of the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement (Moher et al, 2009). This statement requests that a systematic review is a review of a clearly formulated question that uses systematic and explicit methods to identify, select and critically appraise relevant research, and to collect and analyse data from the studies that are included in the review.

Our aim was to summarize the currently reported valid evidence on the association between the use of VWF-containing FVIII concentrates and successful ITI in patients with severe haemophilia A.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

This systematic review is reported in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement (Moher et al, 2009) and was executed according to a protocol written prior to the study.

Search strategy

We performed a comprehensive search using the Medline, EMBASE and Cochrane databases, using the terms: haemophilia, factor VIII, von Willebrand factor, Immune Tolerance, epidemiology and all available FVIII product names.

The full search, which was designed and supervised by an experienced librarian, is listed in the Appendix S1 and was last updated on 3 October 2013. Additional articles were identified from reference lists of the included articles.

Study eligibility criteria

Types of studies

Cohort studies and randomized controlled trials and registries that included at least ten patients and studied the efficacy of ITI for inhibitor eradication were eligible for inclusion.

Types of participants

Severe congenital haemophilia A patients treated with ITI to eradicate inhibitors were considered for inclusion. Studies were eligible with a maximum of 10% non-severe haemophilia A patients of the total cohort.

Interventions

ITI was defined as any treatment with FVIII concentrate, after inhibitor detection (>0·6 Bethesda units [BU]/ml), at least once a week with a minimum of 25 iu/kg/dose. FVIII products were classified as either VWF-containing FVIII concentrates (i.e. plasma-derived FVIII concentrate containing VWF) or as FVIII concentrates devoid of VWF (i.e. plasma-derived FVIII concentrate not containing VWF or recombinant FVIII concentrate).

Outcome measures

The primary outcome was the proportion of patients with successful inhibitor eradication after ITI, defined according to uniform criteria, listed in Table 1.

Table 1. Definitions of outcomes
Primary outcomes Definition
  1. ITI, immune tolerance induction; BU, Bethesda Unit; FVIII, factor VIII.

  2. a

    we specifically designed these definitions for time to success and total of bleeding episodes to ensure a better comparison between studies.

International consensus definitions (Hay & Dimichele, 2012)
Successful outcome ITI/inhibitor eradicationComplete successNegative inhibitor titre (<0·6 BU/ml) (Hay & Dimichele, 2012), FVIII recovery ≥66% of expected, and (good) clinical response to FVIII therapy
Partial successReduction in inhibitor titre <5 BU/ml, with clinical response to FVIII therapy
FailureAny other type of outcome (failure to achieve full or partial success)
Relapse Reappearance of inhibitor after successful eradication by ITI-treatment and inhibitor-free survival of ≤3 months during which at least one confirmed negative Bethesda test (<0·6 BU/ml) (Hay & Dimichele, 2012) has been carried out
Definitions used to enable comparison of the success rates across studies (in Table 2)
Successful outcome definition for systematic reviewLenient definitionNegative inhibitor titre only (<0·6 BU/ml)
Stringent definitionNegative inhibitor titre (<0·6 BU/ml) and FVIII recovery of at least 66% of expected
Time to success Time needed to achieve tolerance, measured from first inhibitor detection until inhibitor eradication (confirmed negative inhibitor titre after ITI)a
Total of bleeding episodes during ITI Total number of bleeding episodes from time of first positive inhibitor test until inhibitor eradication or end of ITI, whichever comes firsta

In order to allow for comparison of success rates between studies we used two additional definitions (Table 2): ‘lenient’ and ‘stringent’. For several studies the outcome could not be re-defined according to these definitions due to unclear description of data or inability to extract necessary data (see Table 2).

Table 2. Success rates according to different definitions
StudyStudy typeSuccess rates ITI
Author (Year)Study definition, n success/totala (%)Lenient definition,b n (%)Stringent definition, n (%)
  1. ITI, immune tolerance induction; SC, single centre; MC, multicentre; R, randomized; RCT, randomized controlled trial; na, not available.

  2. a

    Total = total number of haemophilia A patients with inhibitor receiving ITI with evaluable outcome in study.

  3. b

    If ‘lenient definition’ is not applicable and there is a proportion stated at ‘stringent definition’, than the proportion for the ‘lenient definition’ will be at least as high as the proportion calculated for ‘stringent definition’ but presumably higher.

  4. ‡One patient showed no normalization of factor VIII half-life.

Van Leeuwen et al (1986)SCnanana
Ehrenforth et al (1994)SC14/16 (88)14/16 (88)na
Brackmann et al (1996)SC36/37 (97)37/37 (100)36/37 (97)
Smith et al (1999)SC7/11 (91)nana
Unuvar et al (2000)SC8/14 (57)na8/14 (57)
ter Avest et al (2010)SC18/21 (86)21/21 (100)18/21 (86)
Callaghan et al (2011)SC22/31 (71)na22/31 (71)
Lin et al (2011)SC22/26 (31)22/26 (31)na
Batlle et al (1999)MC9/11 (82)9/11 (82)na
Damiano and Hutter (2000)MC57/81 (70)57/81 (70)na
Rocino et al (2006)MC19/26 (73)21/26 (81)19/26 (73)
Barnes et al (2006)MC17/29 (59)nana
Gringeri et al, 2007)MC9/16 (56)na9/16 (56)
Salviato et al (2007)MC11/16 (69)nana
Unuvar et al (2008)MC5/19 (26)na5/19 (26)
Greninger et al (2008)MC5/11 (45)7/11 (64)5/11 (45)
Kurth et al (2008)MC8/25 (32)na8/25 (32)
Bidlingmaier et al (2011)MC9/14 (64)9/14 (64)na
Kurth et al (2011)MC11/33 (33)na11/33 (33)
Rivard et al (2013)MC22/32 (69)22/32 (69)14/32 (44)
Lenk (2000)R99/126 (79)na99/126 (79)
Haya et al (2001)R26/38 (68)26/38 (68)na
Mariani and Kroner (2001)R128/275 (47)na128/275 (47)
Dimichele and Kroner (2002)R115/164 (70)nana
Coppola et al (2009)R46/87 (53)na46/87 (53)
Hay and Dimichele (2012)RCT37/78 (47)na37/78 (47)

The secondary outcomes were time to success, bleeding rate during ITI, complications of the inhibitor or ITI, and relapse of the inhibitor.

Study selection

Titles and abstracts were screened by two of the authors. We obtained the complete papers of the studies that seemed potentially relevant, from which the final set of studies were selected for inclusion.

Data extraction

Two of the four authors independently extracted data from each study, using a standardized data extraction form. Any disagreement was resolved by consensus or by consulting a third author. When complete overlap of patients across studies was obvious, the most recent update of the registry was included and previous reports were excluded. Partial or suspected overlap is described in the results section.

Assessment of methodological quality

The methodological quality of all studies was judged by two of the authors, using a quality assessment scale derived from the Newcastle-Ottawa Scale. The quality assessment criteria are described in Table S1.

Data synthesis and analyses

As the effect of VWF in ITI is likely to differ according to other determinants of effect of ITI, only those studies that fulfilled both of the following criteria were selected for the meta-analysis:

  1. Intervention: exclusive use of FVIII concentrates for a first cycle of ITI, no other products (e.g. immunomodulants) used for inhibitor eradication treatment.
  2. Outcome: the definition of successful outcome of ITI should at least consist of a negative inhibitor titre.
  3. (<0·6 BU/ml) after ITI on two consecutive tests.

Statistical analysis

The success rate of ITI was calculated as a proportion with a 95% confidence interval (CI).

We used ibm spss statistics (version 20) for Windows (release 20, SPSS Inc., Chicago, IL, USA) for descriptive statistics and Excel for generating plots and calculating proportions and CIs. A random effects model was used to calculate a pooled proportion of success per type of concentrate in the meta-analysis.

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

Selection of articles

The process of study selection is shown in Fig 1. Our literature search yielded 1764 relevant references. After excluding duplicates, 1659 unique references remained. After screening titles and abstracts we excluded 1601 articles. The remaining 58 articles were retrieved in full text for more detailed screening; 25 articles met all inclusion criteria. The other 33 articles were excluded for several different reasons (please see Fig 1).

image

Figure 1. Flow Diagram of study selection. FVIII, factor VIII; ITI, immune tolerance induction.

Download figure to PowerPoint

We included one more article that was identified from reference lists of included studies. Thus 26 articles were included in this systematic review.

Quality of the included studies

The results of the quality assessment of the 26 studies are shown in Fig 2. All studies were found to have methodological limitations. Studies scored particularly poor on the following criteria: definition of study group, definition of outcome and report on financial support.

image

Figure 2. Quality of included studies. Please see Table S1 for details regarding definition of criteria.

Download figure to PowerPoint

image

Figure 3. Proportion of success (with 95% confidence intervals) of studies selected for meta-analysis. *Study no., numbers on y-axis shows study number: see Table 4 for correlating studies. VWF, von Willebrand factor.

Download figure to PowerPoint

Included studies, general

We selected 20 retrospective cohort studies (Van Leeuwen et al, 1986; Ehrenforth et al, 1994; Brackmann et al, 1996; Batlle et al, 1999; Smith et al, 1999; Damiano & Hutter, 2000; Unuvar et al, 2000, 2008; Barnes et al, 2006; Rocino et al, 2006; Gringeri et al, 2007; Salviato et al, 2007; Greninger et al, 2008; Kurth et al, 2008, 2011; ter Avest et al, 2010; Bidlingmaier et al, 2011; Callaghan et al, 2011; Lin et al, 2011; Rivard et al, 2013), five registries (Lenk, 2000; Haya et al, 2001; Mariani & Kroner, 2001; Dimichele & Kroner, 2002; Coppola et al, 2009) and one randomized controlled trial (Hay & Dimichele, 2012). The main characteristics are summarized in Table 3. Eleven studies were conducted in Europe, seven in North-America, one in Asia and seven in multiple continents.

Table 3. Study characteristics
Author (Year)Total patients on ITI (n)aFVIII only as first ITI cycle, n (%)Age (years) at start of ITI, median, (range)Pre-ITI inhibitor titre (BU/ml), median (range)Historical peak inhibitor titre (BU/ml), median (range)Time from first inhibitor detection until start of ITI, monthsi months, (median, days)Time from start of ITI to success (months), median (range)Product used for ITIbStudy type
  1. ITI, immune tolerance induction; BU, Bethesda Unit; FVIII, factor VIII; na, not available.

  2. a

    Not total number of patients included in study, but exclusively the number of severe haemophilia A patients with ITI of which outcome was evaluated.

  3. b

    R = recombinant FVIII concentrate; P-VWF = plasma-derived FVIII concentrate not containing von Willebrand factor (VWF); P + VWF = plasma-derived FVIII concentrate containing VWF; R/P = recombinant FVIII and plasma-derived FVIII concentrates used.

  4. Gringeri et al (2007): all pts with high responding inhibitors and at least one predictive factor of poor response.

  5. Rocino et al (2006); Salviato et al (2007): high responder is classified as titre >5 BU/ml.

  6. Kurth et al (2008): all patients with poor prognosis (>10 BU/ml and 1 extra criterion: e.g. peak >200 BU/ml, age >18 years).

  7. ter Avest et al (2010): all patients age <6 years.

  8. Hay and Dimichele (2012): all good risk patients (age <8 years, pre titre<10 BU/ml, peak titre <200 BU/ml); 115 patients included in study, 78 reached study end-point.

  9. Dimichele and Kroner (2002): titres only described as median of groups “success” and “failure”.

  10. Damiano and Hutter (2000): 81 patients completed ITI, not clearly described how many of these were haemophilia B patients (maximum 6); inhibitor titres shown of total population in dot plot, but individual values not described; 80% of the patients received monoclonal or recombinant FVIII product.

  11. Ehrenforth et al (1994): included 21 inhibitor patients receiving ITI, but exact data only shown for 16 patients [all 16 high responder patients (peak titre >10 BU/ml)].

  12. Salviato et al (2007): all patients receiving ITI were high responders and genotyped.

Van Leeuwen et al (1986)1818 (100)na (1–33)na (0·4–8·7)76 (10–1000)nananaSC
Ehrenforth et al (1994)1616 (100)na41·5 (0·8–1052)124·5 (100–603)na11·3 (2·9–19·7)naSC
Brackmann et al (1996)37nanana (0·8–520)na (3–5500)nananaSC
Smith et al (1999)1111 (100)3 (na)2·9 (0·6–20)14 (9–122)12·1 (364)1·5 (1–3)R/PSC
Unuvar et al (2000)1414 (100)5 (0·2–17)9 (0·7–207)69 (5–1280)30·3 (910)6 (1–8)R/PSC
ter Avest et al (2010)2121 (100)0·75 (0–7)4·5 (1–753)8·3 (1–753)na7 (3·1–14·8)R/PSC
Callaghan et al (2011)3131 (100)5 (0·5–18)4·4 (0–207)34 (0·6–1280)33 (990)2·3 (0·5–34·8)R/PSC
Lin et al (2011)2625 (96)na (0·6–45)na (0·6–700)na (0·5–841)nana (12–60)RSC
Batlle et al (1999)119 (82)4 (0·4–48)2 (0·7–46)33 (0·7–566)na2·5 (0·5–29)RMC
Damiano and Hutter (2000)8154 (67)3 (0·1–44)na (1–1000)na (1–1000)na12·5 (1–113)R/PMC
Rocino et al (2006)2626 (100)4 (1–25)4 (0–60)60 (7–512)na6 (2–39)RMC
Barnes et al (2006)2921 (72)2 (na)8·5 (1–647)na (1–5120)3·5 (103·5)naR/PMC
Gringeri et al (2007)1612 (75)23 (4–54)5 (2–200)54 (16–737)96 (2880)12 (4–30)P + VWFMC
Salviato et al (2007)1616 (100)na4·4 (0–90)300 (30–976)nananaMC
Unuvar et al (2008)1918 (95)9 (1–20)19·2 (3·6–515)80 (6–517)2·5 (75)6 (3–12)P + VWFMC
Greninger et al (2008)117 (64)4 (0·1–36)5 (0–30)89 (2–288)6·5 (195)16·3 (6–28·5)P + VWFMC
Kurth et al (2008)2514 (56)6 (0·1–37)na (0·1–261)140 (2–4833)14 (419)10 (3·0–83·0)R/PMC
Bidlingmaier et al (2011)147 (50)2 (0·7–49)23 (1·4–340)49 (1–564)0·1 (2·4)17 (5·8–29·3)P + VWFMC
Kurth et al (2011)336 (18)2 (0·5–23)22 (0–848)277 (5–2898)0·8 (27)naP + VWFMC
Rivard et al (2013)3232 (100)2 (0·0–5)nana8·7 (260)16·8 (1·2–43·2)RMC
Lenk (2000)126126 (100)na na45 (1–>1000)nananaR
Haya et al (2001)3829 (76)7 (0–57)11 (1–256)67 (6–2984)25 (750)9·9 (2–36)R/PR
Mariani and Kroner (2001)275nana (1–64)7·0 (0–720)54 (1–25 000)17 (510)naR/PR
Dimichele and Kroner (2002)16499 (60)na (0·1–64)nana55·8 (1674)16·3 (0–84)R/PR
Coppola et al (2009)8787 (100)6 (4–59)4·0 (0–200)64 (7–900)na8 (1·5–40)R/PR
Hay and Dimichele (2012)7878 (100)1 (0–8)5·5 (na)36·5 (na)0·2 (5·5)naR/PRCT

A total of 1284 haemophilia A patients with an inhibitor that received ITI were evaluated, ranging from 11 to 275 patients per study. However, there might be partial overlap of patients between the following studies: the International registry (Mariani & Kroner, 2001), German registry (Ehrenforth et al, 1994; Brackmann et al, 1996; Batlle et al, 1999; Lenk, 2000; Bidlingmaier et al, 2011), Spanish registry (Batlle et al, 1999; Haya et al, 2001), North American registry (Damiano & Hutter, 2000; Unuvar et al, 2000; Dimichele & Kroner, 2002; Barnes et al, 2006; Greninger et al, 2008; Kurth et al, 2008, 2011; Callaghan et al, 2011; Hay & Dimichele, 2012) and Italian registry (Rocino et al, 2006; Gringeri et al, 2007; Salviato et al, 2007; Coppola et al, 2009) and separate or national reports of participating centres, two single centre studies (Van Leeuwen et al, 1986; ter Avest et al, 2010).

Five studies reported the effect of ITI exclusively using VWF-containing FVIII concentrate (plasma-derived FVIII products) (Gringeri et al, 2007; Greninger et al, 2008; Unuvar et al, 2008; Bidlingmaier et al, 2011; Kurth et al, 2011), with a proportion of patients included after initial failure of ITI with FVIII concentrates devoid of VWF ranging from 0% to 75%. Four studies (Batlle et al, 1999; Rocino et al, 2006; Lin et al, 2011; Rivard et al, 2013) reported the effect of ITI exclusively using recombinant FVIII concentrate, devoid of VWF. In 12 studies (Smith et al, 1999; Damiano & Hutter, 2000; Unuvar et al, 2000; Haya et al, 2001; Mariani & Kroner, 2001; Dimichele & Kroner, 2002; Barnes et al, 2006; Kurth et al, 2008; Coppola et al, 2009; ter Avest et al, 2010; Callaghan et al, 2011; Hay & Dimichele, 2012) both product types were used and five studies (Van Leeuwen et al, 1986; Ehrenforth et al, 1994; Brackmann et al, 1996; Lenk, 2000; Salviato et al, 2007) did not specify the type of product. The lowest dose regimen used for ITI was 25 iu/kg every other day. In at least nine studies (Batlle et al, 1999; Haya et al, 2001; Mariani & Kroner, 2001; Dimichele & Kroner, 2002; Barnes et al, 2006; Kurth et al, 2008, 2011; Unuvar et al, 2008; Lin et al, 2011) a proportion of the patients (4–37%) received adjunctive treatment for inhibitor eradication such as cyclophosphamide, rituximab and/or prednisolone. The use of adjunctive treatment excluded the subjects from the meta-analysis (13 studies).

Results of individual studies

Primary outcomes

The proportion of successful ITI, according to the studies' separate definitions, ranged from 26% to 97% and is presented in Table 3. The success rates among studies are compared in Table 2, using different definitions for success, as specified in Table 1.

Secondary outcomes

Time to success, as defined in Table 1, could be calculated in 19 studies and the median ranged from 1·5 to 17 months. Fourteen studies reported complications due to the inhibitor and eradication treatment. The range for the incidence of central venous access device (CVAD)-related complications was 0–95% and it was 0–21% for drug-related complications. Bleeding rate during ITI could be calculated in three studies and ranged from 0·0 to 26·4 bleeding episodes per year (Gringeri et al, 2007; Hay & Dimichele, 2012; Rivard et al, 2013). The proportion of patients that had a relapse was reported in nine studies and ranged from 0% to 20%. Due to incomplete reporting of pre-ITI risk factors in the studies, we were not able to group or analyse subjects according to these risk factors.

Data synthesis

Success rate ITI according to product

Thirteen studies were included in the meta-analysis as the remaining 13 studies did not comply with the inclusion criteria for the meta-analysis.

Three studies reported on exclusive use of VWF-containing concentrates (Greninger et al, 2008; Bidlingmaier et al, 2011; Kurth et al, 2011), another three studies exclusively used concentrates devoid of VWF (Batlle et al, 1999; Rocino et al, 2006; Lin et al, 2011) and in the remaining seven studies, both types of concentrates had been used or type of concentrate was not further specified (‘mixed’) (Smith et al, 1999; Lenk, 2000; Unuvar et al, 2000; Barnes et al, 2006; ter Avest et al, 2010; Callaghan et al, 2011; Hay & Dimichele, 2012). We did not have enough information at patient level about product type to make a comparison between the subjects within the mixed studies that used VWF-containing products and those that were treated with products devoid of VWF. The ‘mixed’ studies were taken together and were classified and analysed as a separate group. All studies in the meta-analysis were heterogeneous with respect to the proportion of high- and low-titre patients; hence we specified this as the proportion of patients with successful ITI (Table 4).

Table 4. Separate analysis of selected studies
No.aAuthor, yearTotal number of patients ITIbHigh risk patients, according to pre-titrecn (%)High risk patients, according to peak-titred n (%)Successful ITI, n (proportion)
  1. ITI, immune tolerance induction; na, not available.

  2. a

    Numbers in this column (for studies and study-groups) correspond to numbers on the y-axis in Fig 3.

  3. b

    Number of patients in study that fulfilled all criteria for meta-analysis (Methods).

  4. c

    High risk patient defined as pre-titre higher than or equal to 10 BU.

  5. d

    Peak titre ≥200 BU/ml.

  6. e

    Number shown is a pooled proportion using a random-effects model, calculated from the studies in each specific group.

1 Group: VON WILLEBRAND FACTOR-containing concentrates (studies 2,3,4) 20 11 (55) 6 (30) 13 (0·70) e
2 Greninger et al (2008)72 (29)2 (29)6 (0·86)
3 Bidlingmaier et al (2011)75 (71)1 (14)4 (0·57)
4 Kurth et al (2011)64 (67)3 (50)3 (0·50)
5 Group: non-VON WILLEBRAND FACTOR-containing concentrates (studies 6,7,8) 60 8 (13) 4 (7) 49 (0·84) e
6 Rocino et al (2006)264 (15)2 (8)19 (0·73)
7 Batlle et al (1999)92 (22)1 (11)8 (0·89)
8 Lin et al (2011)252 (8)1 (4)22 (0·88)
9 Group: mixed (both concentrates or concentrate not described) 302 na na 208 (0·74) e
10 Unuvar et al (2000)147 (50)3 (21)8 (0·57)
11 Smith et al (1999)111 (9)na7 (0·64)
12 ter Avest et al (2010)216 (29)2 (10)18 (0·86)
13 Callaghan et al (2011)3111 (35)9 (29)22 (0·71)
14 Hay and Dimichele (2012)780 (0)na37 (0·47)
15 Barnes et al (2006)21nana17 (0·81)
16 Lenk (2000)126nana99 (0·79)

The pooled proportion of successful ITI was 0·70 (95% CI 0·52–0·89) for the patients treated with VWF-containing FVIII concentrates, 0·84 (95% CI 0·75–0·93) for the group treated with non-VWF containing concentrates and 0·74 (95% CI 0·65–0·84) for the ‘mixed’ group.

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

This systematic review summarizes 26 studies on the effect of the type of FVIII concentrate on ITI success in haemophilia A patients. A pooled quantitative meta-analysis of 13 studies restricted to patients receiving their first ITI cycle without immunomodulant co-medication did not demonstrate a higher success rate of VWF containing FVIII concentrates in comparison to FVIII concentrates devoid of VWF.

Despite our careful selection procedure and data synthesis, this comparison needs to be interpreted with care, due to substantial heterogeneity between studies. Populations differed in age and inhibitor titre, ITI regimens, time intervals and definitions of outcome. Due to this heterogeneity of definitions the primary outcome - successful inhibitor eradication after ITI - could only be evaluated by two consecutive negative inhibitor tests because the data on FVIII recovery or half-life were missing. Patients with a lower pre-ITI and peak-titre were more likely to attain successful ITI. To address the mixed populations of low titre and high titre patients, in most studies, we classified the proportion of high-risk patients according to pre-ITI and peak-titre (Table 4). In the meta-analysis the proportion of high-risk patients seemed to be higher in studies reporting ITI with a VWF-containing FVIII compared to the studies reporting on FVIII concentrates devoid of VWF and the ‘mixed’ studies. Therefore, the finding of an apparently lower success rate for treatment with VWF-containing FVIII concentrates (0·70; 95% CI 0·52–0·89) in comparison to treatment with FVIII concentrates devoid of VWF (0·84; 95% CI 0·75–0·93), should be interpreted with care. The relevance of this finding is reduced by the high potential risk of bias that is caused by inclusion of more high risk patients in the VWF-containing concentrate group.

Moreover, in the studies reporting ITI with VWF-containing FVIII concentrates, up to 75% of the patients were treated after initial failure of ITI with a FVIII concentrate devoid of VWF.

In the only randomized trial (Hay & Dimichele, 2012), the success rate was lower than in the observational studies. This may be due to less stringent definitions for outcome used in previous studies and a general trend toward reporting successes. Secondary outcomes were reported in a minority of the 26 studies. Bleeding episodes during ITI, an important clinical outcome, could only be calculated in three studies (Gringeri et al, 2007; Hay & Dimichele, 2012; Rivard et al, 2013).

A randomized controlled study addressing the effect of VWF containing concentrates in ITI is in progress (the RESIST study, http://www.itistudy-resist.com/). This is a well-designed study which includes severe haemophilia A patients with inhibitors at high risk of failure. The study has been ongoing for 5 years, and illustrates once again that it can be very challenging to achieve sufficient patient numbers for randomized trials when studying rare diseases. Due to the difficulties of conducting prospective trials with well-defined and meaningful endpoints in these patients (Dimichele et al, 2012), we need to improve the data quality of observational studies. Studies need to use uniform definitions to avoid heterogeneity. Also, a minimum data set should be defined to standardize the collection of essential data. We described 20 items that should be included in a data set for the evaluation of ITI in haemophilia A (Table S2).

The design of this meta-analyse presents an important concept for future studies that address the same research question. Our framework of inclusion criteria, definitions and minimum data set can be used as a reporting guideline by investigators to enable them to provide a suitable contribution to the present knowledge.

Due to heterogeneity issues, as described above, a regression analyses on baseline risks was not feasible. A suggestion for risk factors that should be taken into account and measured in all future studies: first positive inhibitor titre, pre-ITI inhibitor titre, peak titre and time duration between first positive inhibitor titre and start of ITI.

In conclusion, although limitations due to clinical and methodological heterogeneity hampered firm conclusions, we infer that data synthesis of all relevant valid evidence does not support the idea that VWF-containing products have higher success rates of ITI than non-VWF containing products.

Acknowledgements

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

The authors would like to thank A.E.G. Leenders, clinical librarian at the AMC Medical Library (Amsterdam, the Netherlands) for designing the literature search for this review.

Author contributions

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

A.S. van Velzen has full access to all data, takes responsibility for integrity of the data and accuracy of data analysis. Study design: A.S. van Velzen, M. Peters, J.G. van der Bom. and K. Fijnvandraat. Acquisition of data: A.S. van Velzen, M. Peters, J.G. van der Bom, and K. Fijnvandraat. Analysis and interpretation of data: A.S. van Velzen, M. Peters, J.G. van der Bom and K. Fijnvandraat. Statistical analyses: A.S. van Velzen. Drafting manuscript: A.S. van Velzen and K. Fijnvandraat. Critical revision of manuscript for important intellectual content: J.G. van der Bom, M. Peters, K. Fijnvandraat. Study supervision: J.G. van der Bom, K. Fijnvandraat.

Conflicts of interest

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

A.V. and K.F. have reported receiving unrestricted research support from CSL Behring. K.F. has received unrestricted research funding for various projects from Bayer, CSL Behring and Novo Nordisk. She has received payment for lectures from Bayer and Pfizer and she is a member of the European Haemophilia Treatment and Standardization Board, sponsored by Baxter. J.G. van der Bom has received unrestricted research/educational funding for various projects from Bayer Schering Pharma, Baxter, ZLB Behring, Novo Nordisk and Wyeth. In addition, she has been a consultant to Baxter and Wyeth, and she has been a teacher on educational activities supported by Bayer Schering Pharma. M.P. has received unrestricted funding for various projects from Baxter, Pfizer and Novo Nordisk.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. Author contributions
  8. Conflicts of interest
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
bjh12942-sup-0001-TableS1-S2-AppendixS1.pdfapplication/PDF179K

Table S1. Quality assessment criteria.

Table S2. Minimum data set.

Appendix S1. Full literature search.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.