Previous: Summaries of Findings

Discussion

Clinical aspects
Apart from a few studies (4-6,11) reports on the clinical expression of HAE are mostly in the form of case reports in which one or a few symptoms are in focus. This investigation thus expands the limited number of larger family studies on HAE. By analysing a whole kindred we also hoped to be able to systematise information in order to detect patterns made visible by the relatively high number of affected members.

The clinical studies ran in part in parallel with the mapping of affected family members. We were quite surprised to learn that many of these individuals were deeply grateful to be given an explanation for what had been an enigma to themselves and the medical personnel. As a result, family members were generally interested in participating in the study, and the fact that answers were obtained from every member provided information amenable to reliable statistical evaluation. Sensitivity was achieved by applying an extensive questionnaire as the basis of an interview.

During the exploration of symptoms and inciting events, the affected persons were also given the opportunity to convey information which at first could seem loosely connected with the disease. This was important, as previous family studies have shown that e.g. abdominal symptoms in many cases were not coupled to peripheral swellings, when these in fact were manifestations of the same disease (4). This approach soon paid off, and several original discoveries were made as e.g. the association of eruption of primary teeth with swellings in the oral cavity.

We have also presented symptoms and events which at first did not fit in a model, as it is very unlikely that all details of this disease is known. This information may be of value for future studies, when they can be validated or disputed.

Brain edema
The possibility of a coexisting brain edema is often omitted in case reports and even review articles covering HAE. Less dramatic, but probably of wider relevance is the intriguing idea of a subtle influence of a mild cerebral edema in the beginning of an attack, as mental stress is so often cited by the patients as a triggering factor. It may be that a part of these so-called emotionally induced attacks in fact are a prodrome of the attack itself (45). Headache often precede attacks (5). Later on, the brain may be intimidated by a series of mechanisms during HAE-attacks, including a primary edema, compromised perfusion pressure due to hypovolemia, and probably also by the increased viscosity of the blood due to extravasation of fluid. In cases where the hematocrit reach 75% (47), rheologic disturbances of small vessels in the brain could be a matter of concern.

After respiratory arrest due to laryngeal obstruction and hypoxic brain damage (243), a developing cerebral edema could also be complicated by the liability of edema formation in HAE patients. This is a hitherto unreported phenomenon with therapeutical implications. C1-INH concentrate should be given in this phase (31).

Pulmonary edema
In this thesis pulmonary edema was not observed in any patient., in agreement with statements from a researcher and clinician with thorough knowledge of HAE; " It is interesting to note that the pulmonary tree is never involved in attacks of HAE" (11). Several other authors have, however, observed pulmonary edema in HAE (31,33-36). As these cases with pulmonary edema coincide with laryngeal edema, it is conceivable that the apparent contradiction may be explained by the high negative intrathoracic pressure the patient has to generate to inhale air through a narrow glottis (244). A similar non-cardiogenic pulmonary edema is seen in Intensive Care Units, and is not related to HAE. Following removal of an endotracheal tube which have caused edema near the laryngeal or cricoid cartilages, or because of a subglottic edema in children, large negative intrathoracic pressure can cause transcapillary filtration and edema (245,246). This edema can also follow abrupt relief of a closed or narrow airway, as e.g. when laryngeal spasms suddenly fade or the intubation of the trachea abruptly opens up the airway (247). The latter mechanism may also be highly relevant in HAE patients, whenever intubation or tracheostomy become necessary. The etiology of pulmonary edema following upper airway obstruction represents an interplay between several factors: cardiogenic and neurogenic mechanisms, as well as hypoxia contribute (248).

Edema in infancy
The two clinical studies (I & III) in this thesis both draw attention to the fact that HAE might manifest during the child's first year. In contrast to the message from some comprehensive family studies and case reports (5,6,29), in which e.g. a high number of "colicky babies" were reported as early as in 1966 (4), many reports focus on the adolescence as the period during which HAE manifests. Consequently, few studies are really concerned with the implications HAE might have in childhood. Only recently, a review on the management of HAE in pregnancy in the journal Anaesthesia stated that "HAE is always asymptomatic in infancy" (249). This thesis emphasize even early infancy as a likely age of debut, and encourage physicians dealing with newborns from HAE parents, to suspect an underlying C1-INH deficiency as a potential cause of e.g. unspecific recurrent abdominal pain, diarrhea, upper respiratory disease or skin rash.

Hormonal aspects
This thesis shows that hormonal changes and the intake of estrogen containing medication heavily influence the course of the disease. However, women with two normal C1-INH alleles, may also experience C1-INH reductions. During normal pregnancy C1-INH activity drops (250-253), with a further substantial reduction in preeclampsia (254). These conditions are often accompanied by a marked tendency to edema, during which e.g. a moderate laryngeal edema sometimes hinders intubation of the larynx in the case of cesarean section. The role C1-INH play in the formation of these edematous conditions in normal women not suffering from HAE is unsatisfactorily examined and invites the planning of further studies.

Medication
In the present thesis the adverse effect of estrogen containing pills on HAE, is underlined, including substition in postmenopausal women. Anti-androgen therapy or estrogen to men with HAE, as e.g. when treating prostatic cancer, would most likely lead to aggravation of attacks. Other drugs, possibly contraindicated in HAE include ACE-inhibitors (255,256). The administration of C1-INH concentrate should be concidered to forego the infusion of intravenous contrast media (257), streptokinase (241,258) or t-PA (259), as these agents could readily exhaust the tiny C1-INH pool in HAE patients.

Premature labour
Although case reports show how early uterine contractions, placental abruption or premature labour coincide with abdominal attacks of HAE (96,260,261), systematic evaluation has not previously been published that can corroborate our findings of increased frequency of premature labour in HAE patients. Six of ten women suffering from HAE experienced spontaneous abortions or premature labours in conjunction with symptoms of HAE compared to none of the mothers without HAE (p= 0.037 for number of pregnancies).

In one previous case report, the finding of a dilated cervix was an unexpected discovery, and the cause of referral was the recurrent abdominal cramps the woman usually had during attacks (260). It is therefore possible that other HAE-women suffering from a severe increase of abdominal attacks during pregnancy may in fact have uterine contractions which are masked by the crampy abdominal pain they usually experience during HAE attacks. Even in normal women not suffering from HAE, the recognition of premature labour is difficult, both for the women and the general practitioner. Quite a few women suffering from premature contractions are initially mistaken to have low back pain or abdominal discomfort due to diarrhea or cystitis. In conclusion, the occurrence of premature contractions in pregnant women suffering from abdominal attacks of HAE may be greater than anticipated.

Edema of the uterus may also complicate pregnancy, as a "twenty-pound cannon ball large" uterus has been reported during attacks in a non pregnant women (96).

The mechanisms underlying the start of a normal birth is at present largely unknown. If, however, the generation of bradykinin should prove to be an important physiological mechanism in humans as well as recently shown in rodents (148), very interesting perspectives emerge for a potential use of C1-INH concentrate in the treatment of premature labour in both HAE- and normal women. A preferably recombinant produced C1-INH concentrate could also have advantages compared to the newly developed B2 antagonists, as these oligopeptides cross the placental barrier and may have untoward effects on the fetal kallikrein-kinin system. The 105 kd C1-INH is unlikely to do so (262).

Functional assay of C1 inhibitor

Diagnosis of HAE type II
Determination of C1-INH function is necessary to make the diagnosis of HAE type II, when the presence of a dysfunctional protein from the mutant allele makes the antigenic concentration of C1-INH normal or higher than normal measured by e.g. a Manchini technique. About ¼ of new HAE families have type II HAE. A quantitative functional assay for diagnostic routine purposes did not exist in Norway when the study started. An assay based on a recently developed chromogenic method (263) was therefore established. To take advantage of reagents in a commercial kit kindly provided by Behringwerke AG, the method was modified slightly. The method includes methylamine to prevent the influence of varying alpha-2-macroglobulin concentration. The unpredictable effect of different alpha-2-macroglobulin concentrations has in fact been mentioned as a cause of unreliable C1-INH functional results (264). The method is also easy to perform, and can be carried out in less than 30 minutes. By applying this method in our laboratory, a new family with HAE, the second in the county of Nordland with this disease known so far, was recently found to have HAE type II. The functional assay has also recently been included by the Immunological Laboratory of the National Hospital of Norway

Further employment of C1-INH functional assay
Functional assay of C1-INH may also be applicable in research projects when studying sepsis and conditions with extensive proteolysis like streptokinase infusion. Under these circumstances, C1-INH may be degraded or bound to several proteases which make a gap between the antigenic remnants of the protein and the functional inhibitory capacity (265). Proper function requires an integer, unbound C1-INH protein.

Dilution by citrate
The collection of 9 parts blood into vacutainer tubes containing 1 part citrate solution, results in a dilution of the plasma only, as the isotonic solution stays outside the red blood cells. With a hematocrit of 0.42 the actual protein concentration in citrated plasma would be 84% of that in serum or EDTA-plasma. This bias is accentuated with increasing hematocrit and consequently smaller plasma volume, and all analyses performed in citrated plasma in newborns would heavily depend on the plasma type. In our experience, this is a point which seldom is mentioned when different reference intervals or protein concentrations are compared, and information of additive is in fact regularly omitted in tables citing normal blood chemistry values in the newborn (266). A search on Medline 1989-to date on citrate and dilution yielded no hits.

Effect of heparin
Several functional C1-INH assays are based on exogenously added C1s. As discussed in paper II, heparin can clearly influence these functional assays. Analysis of functional C1-INH values in heparinized plasma may therefore interfere with measurements. It is important to know the effect of in vitro heparin on each functional test, as some authors advocate the use of heparinized samples as the sample of choice for functional analyses of C1-INH (267).

A second point of interest is when samples are obtained from patients having a high content of heparin in vivo. An example is patients undergoing cardiopulmonary bypass or those receiving full-heparinisation. Concentration of heparin during extracorporal circulation may amount to 5-7 IU/ml, and patients treated by full heparinisation 0.2-0.7 IU/ml. This is probably enough heparin to yield falsely high C1-INH functional values in some functional assays.

In a Norwegian thesis heparin was also recently shown to seriously disturb functional analyses of the kallikrein-kinin system, presumably by complexing to prekallikrein to build an enzyme capable of cleaving small chromogenic substrates (268).
Heparin has been shown to inhibit activity of the alternative, classical and terminal pathways of complement by regulating C1, C1 inhibitor, C4 binding protein, C3b, factor H and S-protein (269). Heparin has in fact been used to suppress experimental autoimmune disease in animals (270). Interestingly, inhaled heparin has also been given to HAE patients, apparently to take advantage of the enhancement of heparin on the reaction between C1 and C1-INH (271), although further studies are required to validate the beneficial effects in these two patients.

Cascade activation
All studies of blood samples from HAE patients during attacks suffers from one major approximation; it is not known to what extent local tissue concentrations of C1-INH and the proteases are reflected in the circulation. In the edematous regions, which generally are well delimited, it is conceivable that the protease/inhibitor ratio in fact is quite different from that in the circulation, and that samples from the circulating blood at best can give clues to the ongoing reactions in the tissues. This is further illustrated by the lack of correlation between serum values of C1-INH and attacks (272). The fact that approximately 70% of the C1-INH pool is located extravascularly, also points to a crucial role of this compartment (273). Future studies should analyse specimens from edematous tissue if possible, and the examination of lymphatic fluid, especially when draining edematous regions could also yield important information as to the tissue cascade balance.

A possible influence of hemoconcentration on results
Attacks of angioedema have been shown to be accompanied by serious hemoconcentrations in some cases (4). Unfortunately, hematocrit measurements were not obtained throughout our attack series.

To our knowledge no reports have actually dealt with the problem of comparing remission samples with normal hematocrit to attack samples with potentially severely decreased plasma volume. There are several interesting aspects regarding hemoconcentration;
Firstly, the use of citrated plasma could result in a higher dilution in the attack samples with low plasma volume, compared to remission samples, and thus give falsely low values (see previous discussion).

The second question is how the loss of fluid from the vascular compartment affects protein concentration in the remaining plasma. Hemoconcentration of red blood cells does not necessarily mean plasma-protein concentration. How the passage of plasma proteins during attacks of HAE depends on molecular weight of the proteins is a very crucial question, but to our knowledge largely unknown. Since the edema is non-pitting the protein concentration can be assumed to be higher than if only water were let out, and the demonstration of wide endothelial gaps allowing even small amounts of platelets to cross, suggest that nearly all plasma proteins can escape in areas of HAE attacks (88,274,275). If the endothelial gaps observed during attacks let all proteins pass freely, the protein concentration would be expected to be almost the same inside the vessel as outside. If only water and proteins with a small MW could pass then the MW of the actual protein investigated should be of interest before making any corrections for the increase in hematocrit.

As this issue was raised by a referee, we examined our samples in the attack-study for the concentration of 4 plasma proteins with different MW, namely Albumin, IgG, IgA and IgM.

We could not detect differences between the attack and remission samples with respect to the protein concentration of these proteins. The findings coincide with those found when examining the same attack and remission samples with analyses of the cascade systems; activation products were markedly increased while no difference in native proteins between attack and remission samples were observed. This pilot investigation of how proteins with different molecular weight cross the endothelium in HAE, suggests that proteins with a wide MW move freely through the large gaps in endothelial cells that emerge during attacks of angioedema. Due to lack of space, these results were not included in paper V.

Coagulation is moderately enhanced in HAE
A modest enhancement of fibrinolysis has been shown in HAE patients as previously mentioned and also discussed in papers V and VI, but a tendency of bleeding is absent. By using a sensitive assay, this work revealed a very modest increase of thrombin during attacks. This could indicate that the slight increase in fibrinolysis could be balanced by a correspondingly enhanced coagulation. Our finding of a moderate increase in activated FVII during remission may be one mechanism by which thrombin is generated. This pathway is believed to be via an increase of FXIIa in HAE patients that directly activates FVII. Previous studies, however, have failed to directly demonstrate an increase of FXIIa in HAE. Our analyses of FXIIa/C1-INH complexes did not detect a significant increase either. Very recently though, Agostoni and colleagues presented FXIIa activation in HAE, measured by a new and very sensitive monoclonal antibody (207), (Kinin 95, C55).

Newly, FXI was demonstrated to have C1-INH as one of its major inhibitors (215). We could not demonstrate an increase in FXIa activating during attacks. This may be explained by the new concept of coagulation, where FXI does not participate in the initiating steps, but participates later in the coagulation process to sustain the process (208). The minute amounts of thrombin formed in HAE patients may not have been able to activate FXI.

The role of a possible functional protein S deficiency in HAE is a very interesting debate, but one which has barely started, and theories for a mechanism are lacking (219,220). The question of whether C4b split products in HAE interferes with the binding of protein S by the common binding protein, C4BP, represent a challenging future research area. Of interest in this aspect is the fact that serum amyloid P component (SAP), a normal constituent of blood and extravascular tissues, inhibits C4BP function, even though SAP and C4b bind to distinct sites on C4BP (276). Recent reports of activated protein C resistance interfering with functional protein S assays must be kept in mind, as this can give spuriously low protein S function (277).

Complement activation throughout the cascade in HAE
Both the paper on physical stress and the study of attacks showed for the first time directly a modest activation of complement beyond C2. A mild activation of C3 has previously been suggested by the indirect observation of case reports of increased radiolabelled C3 turnover. Ruddy et al. apparently found increased C3 turnover in 4 HAE patients, but controls were patients with kidney disease (278). Later the same values were compared to normals in another study, but statistics were not performed (279). The frequently cited study by Alper and Rosen (280) also used this method, and found C3-turnover normal during remission in one HAE patient (2.98%), and slightly increased (3.62%) in another during an attack, and concluded that "No important abnormalities of metabolism were found in hereditary angioedema".

The finding of C3 and even terminal complement activation implicates that in addition to bradykinin, which presumably is the dominating mediator, the liberation of the potent anaphylatoxins C3a and C5a may contribute to the edema. C4a, which is liberated in advance of C3a and C5a in the cascade, but in higher amounts in HAE, is far less potent than C3a or C5a, and would therefore be expected to play a minor role (94,281).

The anaphylatoxins liberate histamines which participate in edema formation. In the ventricular mucosa histamine is likely to increase acid secretion in HAE patients. This may explain the concurrence of ulcer and heartburn with attacks and why some reports have shown increased level of histamine in urine (14,47). Bradykinin may actually also stimulate histamine release (88) but the relative contribution of these mediators in liberating histamine in HAE is unknown. C3a and C5a are also potent chemotactic factors, and leucocytosis is in fact sometimes observed during attacks (282). Finally, this observation also lends support to the findings by a few authors of some beneficial effect of antihistamines in HAE patients (8). The effect of antihistamines and steroids is limited, however. They are insufficient as acute therapy and their use has not prevented fatal outcomes (283).

Applicability of HAE-studies to edema in non-HAE patients
Hereditary angioedema is an important example of the so-called "experiments of nature", which include disorders that are characterized by a well-defined biochemical defect that provide an opportunity to define basic pathophysiological mechanisms. In addition to those previously mentioned in this thesis, several reports now show that the pathophysiological mechanisms studied in HAE patients may apply to other persons in extreme situations, although they originally have normal C1-INH production.

Clinical applicability of cascade blockers, bradykinin antagonists or C1-INH concentrate
Cascade activation obviously play a role in cerebral pathology in non-HAE patients as well. Nafamostat mesilate (Fut -175), a potent synthetic inhibitor of the complement and kallikrein-kinin systems, reduces focal or diffuse cerebral ischemia in the acute stage after subarachnoid hemorrhage in humans (284). As previously discussed, bradykinin is proposed to play a key role in brain edema after cold lesion, concussive brain injury, traumatic spinal cord and ischemic brain injury, and by administering bradykinin antagonists, advantageous effects on brain edema in humans suffering from traumatic brain injury have lately been presented (KININ 95, L56,L58, L59). Bradykinin antagonists have recently been approved for clinical trials, and makes it possible for the first time to selectively block the effect of bradykinin on B-2, the most important bradykinin receptor (157,285-287).

According to the formerly mentioned role of bradykinin, it would now be attractive to employ bradykinin blockers against snake venoms, acute pancreatitis, septic states, the capillary leakage following major operations or cardiopulmonary bypass, and streptokinase infusion. Some clinical trials are already running. These and the suggestions of other exiting trials mentioned above will eventually outline the importance of bradykinin in edematous states also in humans without HAE.

Many of these beneficial effects could possibly also be achieved by using high doses of C1-INH concentrate, which will have the advantage of both blocking the production of bradykinin and limiting classical complement activation (288-298). It would be an advantage if C1-INH could be produced by recombinant techniques, thereby providing generous and affordable amounts for both virus-safe HAE substitution therapy and research purposes. The potential for a therapeutical use of C1-INH concentrate is a very challenging issue, and reaches far beyond substituting HAE patients.

Bradykinin antagonists may also have therapeutical advantages in HAE patients, as it is not derived from human plasma, and the molecule is so small that administration via routes other than the intravenous may succeed.

Conclusions
In addition to a broad presentation of clinical symptoms and triggering factors, original observations demonstrate the effect of estrogen substitution as a cause of flares of angioedema and suggest a role for HAE in provoking premature labour. Observations were based upon personal interviews and extensive questionnaires, with a 100% response rate, in a large family with 63 members.
Guidelines for the interpretations of a new and modified functional assay for C1-INH are presented, along with original reference intervals for newborns. The significance of making the diagnosis of HAE at birth is demonstrated.

For the first time all cascades are studied simultaneously in HAE patients, utilizing several new sensitive monoclonal antibodies not previously applied on this condition, and the study confirms that the disease seriously influence the kallikrein-kinin and the complement system. In the latter an activation beyond C3, throughout the terminal part is presented for the first time in HAE, suggesting a role for the anaphylatoxins in mediating symptoms. A modest increase in fibrinolysis during attack is confirmed and new observations of higher FVIIa levels in HAE patients and a modest thrombin production during attacks, suggest a minor involvement of coagulation as well.

Finally, the study of the mechanisms underlying the capillary leakage in hereditary angioedema suggest a potential use of C1-INH concentrate also in the treatment of several disorders in humans not suffering from HAE.

Next: References