Edited by: Bodo Niggemann
Target levels of functional C1-inhibitor in hereditary angioedema
Article first published online: 19 SEP 2011
© 2011 John Wiley & Sons A/S
Volume 67, Issue 1, pages 123–130, January 2012
How to Cite
Hack, C. E., Relan, A., van Amersfoort, E. S. and Cicardi, M. (2012), Target levels of functional C1-inhibitor in hereditary angioedema. Allergy, 67: 123–130. doi: 10.1111/j.1398-9995.2011.02716.x
- Issue published online: 12 DEC 2011
- Article first published online: 19 SEP 2011
- Accepted for publication 22 August 2011
- angioedema attack;
- hereditary angioedema;
- human C1-inhibitor
To cite this article: Hack CE, Relan A, van Amersfoort ES, Cicardi M. Target levels of functional C1-inhibitor in hereditary angioedema. Allergy 2012; 67: 123–130.
Background: Hereditary angioedema (HAE) is a heterozygous deficiency of first component of complement-inhibitor (C1INH). Insufficient C1INH activity leads to uncontrolled activation of plasma cascade systems, which results in acute angioedema attacks in patients with HAE. Plasma-derived or recombinant C1INH products are approved for the treatment of such angioedema attacks. The target level of C1INH activity needed to achieve optimal efficacy, however, remains unknown. We determined the plasma level of C1INH associated with optimal clinical efficacy in the treatment of angioedema attacks.
Methods: Efficacy and pharmacokinetic data were reviewed from recently published placebo-controlled randomized trials in the treatment of HAE with either plasma-derived or recombinant C1INH products, tested at various doses.
Results: A dose-dependent effect was observed on time to the beginning of relief of symptoms, on time to resolution of symptoms, and on the response rate within 4 h. Optimal efficacy of C1INH therapy is achieved at doses ≥50 U/kg. This dose increases plasma C1INH activity in almost all patients to values ≥0.7 U/ml (70% of normal), the lower limit of the normal range. The differences in half-lives of the various C1INH products do not have an obvious effect on clinical efficacy.
Conclusion: A review of the efficacy and pharmacokinetic data from recently published controlled studies in the treatment of HAE attacks suggests that efficacy of C1INH therapy is optimal when C1INH activity levels are restored to the normal range.
first component of complement
subcomponent of C1
subcomponent of C1
mannan-binding lectin-associated protease
recombinant human C1-inhibitor
Hereditary angioedema (HAE) resulting from a heterozygous deficiency of the plasma protein first component of complement-inhibitor (C1INH) has a prevalence of one in 50 000 (1, 2). The generation of vasoactive peptides such as bradykinin, resulting from insufficient control of the contact system, is a main step in the pathogenesis of angioedema in HAE among others because during attacks, but not in basal conditions, circulating levels of bradykinin and cleaved high molecular weight kininogen are increased (3–5), while levels of contact system proteins are decreased (6). Moreover, gain-in-function mutations of factor XII give rise to angioedema in the absence of C1INH deficiency (7, 8). Third, mutation of C1INH yielding deficient inhibitory activity toward complement proteases, but not toward the contact system proteases, does not result in HAE (9). Finally, drugs that block the activity of bradykinin or kallikrein reduce the intensity and duration of an HAE attack (10, 11).
C1-inhibitor interacts with the active site of activated C1 to form stable, covalent complexes, which are cleared by the liver (12). C1INH also directly interacts with native C1 to prevent the autoactivation of C1 (13, 14), thereby inhibiting its own consumption. The threshold level of C1INH to prevent the autoactivation of C1 has been claimed to be ∼35% of normal levels or less (13, 14). Most patients with HAE have functional C1INH below this critical level (15). Therefore, autoactivation and subsequent consumption of C1INH have been proposed as an explanation for the lower than expected levels of functional C1INH in the heterozygous HAE patients (∼20% instead of 50% of normal).
Administration of exogenous C1INH has been used for decades to treat acute angioedema attacks with the rationale that increasing functional C1INH levels in the circulation of patients with HAE restores normal homeostatic control of plasma cascade systems and stops the formation of vasoactive peptides that mediate angioedema. Four double-blind placebo-controlled trials have been performed with three different plasma-derived C1-inhibitor (pdC1INH) (16–18) and one recombinant human C1INH (rhC1INH) product (19) in symptomatic HAE patients. As extensive dose-finding studies are not feasible in symptomatic HAE because of the low prevalence of the disease and the heterogeneity of the attacks, the results from these published double-blind placebo-controlled trials were compared to estimate the target levels of C1INH activity for optimal clinical efficacy.
Three pdC1INH products, Cetor/Cinryze® (Sanquin, Amsterdam, the Netherlands), Berinert P® (CSL-Behringwerke, Marburg, Germany), and C1-inactivator (Immuno-Baxter, Vienna, Austria, withdrawn from the market in 2003) (16–18), as well as one rhC1INH (Ruconest® or Rhucin®; Pharming, Leiden, the Netherlands) (19), have recently been tested in randomized placebo-controlled studies for efficacy and safety in symptomatic HAE patients (summarized in Table 1). RhC1INH has the same inhibitory properties as pdC1INH (20) and has recently received market authorization for the treatment of acute angioedema attacks in Europe. There is no objective single clinical or biochemical parameter that reflects the severity and course of acute angioedema attacks at various anatomical locations. In the C1INH intervention studies, efficacy has therefore been measured using patient-reported outcome measures. Because such measures in general are prone to bias, open label data are not considered in the present analysis and only efficacy data of the randomized double-blind placebo-controlled studies have been used for the present analysis.
|Study||Inclusion criteria||Dose||N||Primary endpoint||Secondary endpoint|
|Craig et al. (17)||Moderate, severe||Placebo||42||Time to onset of symptom relief||Time to complete resolution|
|Abdominal, facial||10 U/kg||39||Number of vomiting episodes|
|Onset <5 h||20 U/kg||43||Worsening between 2 and 4 h|
|Zuraw et al. (18)||Moderate, severe||Placebo||33||Time to unequivocal relief of symptoms||Response rate <4 h|
|Abdominal, facial, external genitalia, no laryngeal||1000 U/2000 U||35||Time to complete resolution|
|Onset <4 h|
|Kunschak et al. (16)*||All types of attacks, not life-threatening||Albumen||11||Time to onset of relief||Time to resolution of symptoms|
|Onset <5 h||25 U/kg||11|
|Zuraw et al. (19)||Moderate, severe (VAS)||Saline||29||Time to onset of relief by VAS||Time to minimal symptoms by VAS|
|All types of attacks, not life-threatening||50 U/kg||12|
|Onset <5 h||100 U/kg||29|
Inclusion criteria differed slightly among the trials with the C1INH products. All trials evaluated time to patient-reported onset of relief of symptoms as primary efficacy outcome, although the method to measure onset of relief differed (Table 1). The studies also had comparable secondary efficacy outcome, i.e. time to resolution of symptoms or to minimal symptoms. The dose of C1INH administered in the trials varies from 10 to 100 U/kg (1 U of C1INH activity is the amount of C1INH activity present in 1 ml of pooled normal human plasma). To better compare the outcomes of the C1INH trials by dose group, the outcome of the dose groups was corrected for the outcome of the corresponding placebo group. We did this by calculating the time to onset of relief of symptoms for the C1INH dose groups as percentage reduction compared with the time to relief observed in the corresponding placebo group. In addition, response rates within 4 h corrected for the response in the corresponding placebo group were calculated for the C1INH dose groups by setting the response in the corresponding placebo group at 0%, whereas 100% was used when all patients in a group responded within 4 h.
Threshold level of C1INH activity in HAE to prevent angioedema attacks
Patients with HAE have levels of functional C1INH ranging from <20% to 50% of normal (<0.2–0.5 U/ml). Based on the comparison of levels of C1INH activity in HAE patients with and without attacks, a threshold baseline level of 38% of normal (0.38 U/ml) was proposed to protect the occurrence of attacks (21). Review of unpublished data on functional C1INH levels of 277 attacks in 115 patients with HAE treated with rhC1INH revealed three attacks (1% of all attacks evaluated) in three patients with baseline functional C1INH >0.38 U/ml, the highest level being 0.46 U/ml. Analytical errors in the laboratory cannot fully explain these relatively high levels of functional C1INH because one of these patients, with a HAE type 1, also had a functional C1INH level >0.38 U/ml during screening as well (0.42 U/ml at screening and 0.39 U/ml at presentation to the clinical center), whereas the other two patients had a level of <0.28 U/ml at screening, when patients were asymptomatic. Conversely at screening, three patients were identified with a C1INH activity level >0.38 U/ml, the highest level being 0.6 U/ml. Among these three patients was the one who also had a C1INH level >0.38 U/ml when presenting with an acute attack. The other two patients had a C1INH activity level >0.38 U/ml during screening but had <0.28 U/ml when presenting with an attack. This fits well with the experience of one of the authors (MC) who, among 591 asymptomatic patients referred to him, identified 15 patients with C1INH values ≥0.38 U/ml but below 0.5 U/ml, whereas 348 patients had values <0.2 U/ml (M. Cicardi, unpublished data). These 15 patients all had family history of angioedema caused by hereditary C1INH deficiency. Based on antigenic C1INH levels, two were diagnosed to have HAE type 2 and 13 HAE type 1. Thus, occasional patients may have C1INH activity >0.38 U/ml when presenting with an attack.
The published pharmacokinetics of the C1INH products tested in randomized placebo-controlled trials in HAE is summarized in Table 2. The relatively short plasma half-life of rhC1INH results from clearance via the asialoglycoprotein and mannose receptors owing to incomplete capping of glycans of rhC1INH with sialic acids and the presence of oligomannose-type N-glycans, respectively (22–24). The plasma half-life of pdC1INH is longer than that of rhC1INH and varies considerably in patients with HAE (see Table 2). Observed half-life of the same pdC1INH product may vary from 10 to 96 h (25). This variability of half-life of pdC1INH products probably reflects inter-individual differences in baseline C1INH levels, as well as levels of activated target proteases for C1INH.
|Product||References||Manufacturer||Dose intravenous||Half-life (h)|
|Berinert P||Kreuz et al. (25)||CSL-Behringwerke||1000 U||33 ± 20*|
|Martinez-Saguer et al. (30)||1000 U||41 ± 23|
|Bernstein et al. (31)||10–20 U/kg||32.7 (17–49)†|
|Cinryze/Cetor||Cocchio et al. (26)||Viropharma/Sanquin||1000 U||56 ± 36|
|2000 U||62 ± 38|
|C1-inactivator||Kunschak et al. (16)||Immuno-Baxter||25 U/kg||38 ± 20|
|Ruconest||Van Doorn et al. (22)||Pharming||100 U/kg||2.9 ± 0.6|
|50 U/kg||1.6 ± 0.14|
Efficacy of C1INH products to treat acute angioedema attacks.
The published efficacy outcomes of the four randomized double-blind placebo-controlled trials are summarized in Table 3. All studies with doses of >10 U/kg met their primary and secondary endpoints, except for the study by Kunschak et al. (16). The latter study evaluating pdC1INH at 25 U/kg vs placebo met its primary endpoint, time to relief of symptoms (P = 0.007), but missed its secondary endpoint, time to resolution of symptoms (P = 0.09). Notably, the limited number of patients (11 in each treatment group) may have accounted for this. Hence, the efficacy outcomes of the randomized placebo-controlled studies provide compelling evidence that C1INH administration at doses >10 U/kg shortens the times to onset of relief of symptoms and to the resolution of symptoms as compared to treatment with placebo.
|Study||Dose group||Primary endpoint outcome (min)||Secondary endpoint outcome (min)||Response rate within 4 h (%)|
|Craig et al. (17)||Placebo||90||467||59†|
|20 U/kg||30 (P = 0.0025)‡||295 (P = 0.0237)‡||85|
|Zuraw et al. (18)||Placebo||>240||1500||42|
|1000 U/2000 U||120 (P = 0.02)||738 (P = 0.004)||60 (P = 0.06)|
|Kunschak et al.* (16)||Albumen||1020||1560||27|
|25 U/kg||50 (P = 0.007)§||844 (P = 0.09)§||82|
|Zuraw et al. (19)||Saline||495||1210||41|
|50 U/kg||122 (P = 0.013)¶||247 (P = 0.001)¶||100 (P = 0.056)|
|100 U/kg||66 (P < 0.001)¶||266 (P < 0.001)¶||93 (P < 0.001)|
Target level of C1INH activity for efficacy
Appropriate dose-finding studies of C1INH in symptomatic HAE patients have not been executed. However, doses of 10–100 U/kg of various C1INH preparations have been evaluated in the randomized placebo-controlled studies. A comparison by dose is given in Table 4, which is based on the published data of these studies. In one study (18), a fixed dose of 1000 U per patient was used, with the provision for a second dose of 1000 U after 1 h in case of an unsatisfactory clinical response. Although 23 of the 35 patients in the C1INH arm indeed received a second dose, for uniformity and because this second dose was given after a delay of 1 h, in Table 4 this dose was set at 13 U/kg, assuming a single dose of 1000 U of C1INH and a body weight of 80 kg, which was the mean body weight of patients in the C1INH group of that study (18). The median times to onset of relief of symptoms of the individual studies (Table 4, second column) were shorter in the C1INH dose groups than any of the times observed in the placebo groups, except for the placebo group of the study by Craig et al. (17) (Table 4). The relatively short time to beginning of relief in the placebo group of that study (90 min) apparently was attributed to inclusion of patients with mild or moderate attacks because this time was considerably longer (810 min) in patients with severe attacks in the placebo arm (17).
|Dose group||Data from reference||Time to relief (min)||Time to onset of relief as percentage of placebo time to onset of relief*||Time to resolution (min)||Time to resolution as percentage of placebo time to resolution||Response rate (%)||Response rate minus response rate in placebo group (%)†|
|Placebo (10; 20)‡||Craig et al. (17)||90||100||467||100||59||0|
|Placebo (13)||Zuraw et al. (18)||240§||100||1500||100||42||0|
|Placebo (25)||Kunschak et al. (16)||1020||100||1560||100||27||0|
|Placebo (50; 100)||Zuraw et al. (19)||495||100||1210||100||41||0|
|10 U/kg||Craig et al. (17)||70||78||1200||>100||70||11|
|13 U/kg¶||Zuraw et al. (18)||120||50||738||49||60||18|
|20 U/kg||Craig et al. (17)||30||33||295||63||85||26|
|25 U/kg||Kunschak et al. (16)||50||5||844||54||82||55|
|50 U/kg||Zuraw et al. (19)||122||25||247||20||100||59|
|100 U/kg||Zuraw et al. (19)||66||13||266||22||93||52|
The median time to the beginning of relief of symptoms in the patients treated with 50 U/kg rhC1INH was nearly twice as long as in patients treated with 100 U/kg (122 vs 66 min, respectively, Table 4). However, the 95% confidence intervals were overlapping, and the Kaplan–Meier plots of both dose groups showed no difference in response, suggesting comparable efficacy of either dose (19). Indeed, the secondary endpoint outcome and the response rate supported comparable efficacy of 50 and 100 U/kg rhC1INH (Table 4).
C1-inhibitor treatment dose-dependently reduced the time to onset of relief of symptoms in the symptomatic HAE patients with the largest reduction in the patients treated with pdC1INH at 25 U/kg (Fig. 1). However, the placebo group of this C1INH dose group had the longest time to onset of relief of symptoms among all placebo groups (Table 4), which influenced the reduction in the time to onset of relief of symptoms of this particular C1INH dose group. Analysis of resolution of symptoms in the same way revealed that pdC1INH at doses of 20 or 25 U/kg reduced time to resolution by approximately 50%, whereas rhC1INH at 50 or 100 U/kg reduced time to resolution by about 80% (Table 4). Corrected response rates within 4 h also showed a dose-dependent increase with optimal responses at 50 and 100 U/kg (Fig. 2).
A target level of C1INH activity for optimal therapeutic efficacy of C1INH treatment in symptomatic HAE is unknown. The results of the analysis reported here support that for optimal efficacy, C1INH activity levels should be restored to the normal range.
A threshold baseline level of C1INH activity in the circulation of 38% (0.38 U/ml) of normal has been claimed to protect against the development of attacks (21). We identified three attacks with baseline functional C1INH level >0.38 U/ml, the highest level being 0.46 U/ml. Thus, the target level of C1INH activity for protection against the development of attacks may be >0.38 U/ml, the more because the C1INH levels just mentioned were taken after the attack already had started and presumably had consumed some C1INH. Therefore, the baseline levels prior to the attack may have been somewhat higher. Indeed, prophylaxis with pdC1INH (18) shows that administration of 1000 U of pdC1INH every 3–4 days reduces attack rate but cannot prevent all attacks, despite maximum plasma concentrations (Cmax) of 0.68 U/ml (26).
Analysis of the efficacy data of the randomized controlled studies supports that optimal efficacy occurs at a C1INH dose of 50 U/kg, regarding both time to resolution of symptoms and response rate within 4 h. However, for the time to the beginning of relief, doses higher than 25 U/kg do not seem to be more efficacious. To correct for differences in response time in the placebo group, the percentage reduction in the median time to onset of relief in the C1INH groups compared with that of the corresponding placebo group was calculated (Fig. 1). In the study with 25 U of pdC1INH per kilogram (17), this time was relatively long (>1000 min) in the placebo group, possibly due to the limited number (11) of patients tested. This may have led to overestimation of the corrected efficacy of the 25 U/kg dose.
The relationship between dose and clinical efficacy (Table 4; Figs 1 and 2) suggests that there is no fundamental difference between the C1INH products regarding efficacy, despite their different half-lives (ranging from 1.6 h for rhC1INH at 50 U/kg to >30 h for pdC1INH products, Table 2). Notably, relapses were not seen with any of the C1INH products, including rhC1INH that has a short half-life (22). Apparently, for optimal clinical efficacy of C1INH, peak plasma levels are more relevant than half-life. In other words, for clinical efficacy for acute attacks, only a short-lasting temporary increase in C1INH activity levels is needed. This raises intriguing questions regarding the mechanism of action of C1INH as a treatment for acute angioedema attacks. First, one may hypothesize that for clinical efficacy in the treatment of acute attacks, only a short increase in C1INH peak plasma level is required to inactivate a key protease, possibly factor XIIa, which once inactivated by C1INH cannot become activated again. Alternatively, one may postulate that clearance of C1INH from the circulation also reflects tissue absorption. Estimation of plasma levels does not allow us to distinguish whether the infused C1INH is cleared by the liver and is catabolized or whether it binds to yet unidentified receptors on endothelial or other cells to exert its physiological role, nor is it obvious at what rate the bound C1INH is eliminated. For example, pdC1INH may contain sialyl Lewis-X (27). In addition, it is not clear whether pdC1INH has similar properties in this respect as rhC1INH. rhC1INH is considerably more effective than pdC1INH as a treatment for stroke in a mouse model (28), which may point to differences in binding between rhC1INH and pdC1INH to cells of involved organs. It could be hypothesized that absorption of rhC1INH and pdC1INH into tissues may result in longer bioavailability than suggested by the plasma levels, which would support the prophylactic use of C1INH even when the plasma half-life is short.
Population pharmacokinetic modeling of the results with rhC1INH (manuscript submitted) indicates that a dose of 50 U/kg will bring C1INH activity levels above the lower limit of normal (0.7 U/ml) in virtually all patients. As this dose seems to be optimal for efficacy, the pooled data of all randomized placebo-controlled trials with rhC1INH in symptomatic HAE patients suggest that for optimal efficacy, circulating C1INH activity levels should be in the normal range. Observations with rhC1INH in asymptomatic HAE patients have revealed that levels of functional C1INH below 0.7 U/ml coincide with increasing activation of C4 (22), indicating that C1INH levels need to be in the normal range for optimal control of complement activation. This likely also holds for contact activation because administration of pdC1INH at a dose that yields a mean Cmax of approximately 0.5 U/ml in asymptomatic HAE patients did not fully restore the increased plasma levels of cleaved high molecular weight kininogen to the normal range (32). Furthermore, other parameters of the contact system such as the increased factor XIIa and kallikrein levels do not normalize upon infusion of C1INH at a dose, 1000 or 2 × 1000 U/kg, that does not bring C1INH activity levels to the normal range (33). Thus, these biochemical data support that in symptomatic HAE patients, C1INH activity levels should be restored to >0.7 U/ml for optimal clinical efficacy as well as for optimal inhibition of the plasma cascade systems.
All randomized controlled trials evaluating C1INH therapy in symptomatic HAE patients have met their primary endpoint, implying that even with suboptimal doses of C1INH, a significant proportion of the patients respond to therapy. Future research should reveal whether patients with HAE responding to suboptimal C1INH dose by variation have the highest C1INH activity levels upon infusion with C1INH, or whether in these patients, contact activation is already controlled at lower C1INH activity levels. In conclusion, analysis of the pooled data from randomized placebo-controlled trials with different C1INH products suggests that for optimal clinical efficacy in symptomatic HAE as well as for optimal control of activation of the complement cascade, peak plasma C1INH levels need to be restored within the normal range.
Each author has contributed to the design and outcome of the analysis described in this article.
Conflict of interest
Drs AR and ESVA are employees of Pharming Technologies BV. CEH and MC reported to have received consultancy fees from Pharming Technologies BV.
- 20Ruconest summary of product characteristics (SPC). Available at: http://www.ema.europa.eu/docs/en_GB/document_library/EPAR_-_Product_Information/human/001223/WC500098542.pdf.
- 26Cinryze, a human plasma-derived C1 esterase inhibitor for prophylaxis of hereditary angioedema. P T 2009;34:293–295., .