Monitoring recurrent angioedema: Findings from the Turkish angioedema control test validation study

Abstract Background Determination of control level in recurrent angioedema (RAE) is necessary to guide management. Here, we validated a Turkish version of the angioedema control test (AECT) for 4‐week (AECT‐4wk) and for 3‐month (AECT‐3mth) and assessed their utility in monitoring RAE. Method The recommended structured translation process for patient‐reported outcome measures was completed. The final versions were administered to 51 patients with mast cell‐mediated angioedema (MMAE) and 38 patients with hereditary angioedema, and the minimal clinically important difference (MCID) was determined. Additionally, anchor surveys comprising angioedema activity score for 28 days (AAS‐28 day), visual analog score for angioedema control, Likert scale for the control level from the patient's perspective (LS‐AEC), angioedema quality of life, short form‐12 (SF‐12) and patients' assessment of treatment sufficiency were applied. Results The Turkish AECT versions showed good convergent validity with a substantial correlation with anchor tools and known‐group validity. Excellent internal consistency and reproducibility were observed. Equal or more than 10 of 16 points scored with the AECT‐4wk and AECT‐3mth identified patients with well‐controlled disease. The disease activity, control and burden parameters were consistent with the disease control level defined depending on the cut‐off point 10 of AECT. Three‐point changes in AECT‐4wk and ‐3 mt could detect MCID in disease control in all patients. Conclusions Turkish AECT versions are valid and reliable tools for assessing and monitoring disease control in patients with RAE. The use of the Turkish versions of the AECT in routine patient care, clinical trials and angioedema research is recommended.

treatment of acute attacks and the prevention of attacks with prophylactic therapies, which differ depending on the underlying RAE pathomechanism. 2,3dependent of the RAE type, the aims of treatment are complete disease control, the absence of signs and symptoms, and a normal quality of life.Therefore, determining the disease control levels with a standardized, validated, simple and reliable tool is essential.International guidelines recommend the use of patientreported outcome measures (PROMs) to evaluate disease activity, disease control, and impairment of the quality of life in RAE. 2,3This includes the angioedema control test (AECT), which was developed and validated as a retrospective instrument to measure the disease control level in patients with RAE. 9,10ere are two versions of the AECT, where one measures disease control during the last 4 weeks (AECT-4wk) the other one during the last 3 months (AECT-3mth).Both AECT versions comprise four questions with five answer options each (scored with 0-4 points).
Accordingly, the total AECT score ranged from 0 to 16 points.The higher the score, the higher the level of disease control.A total score <10 points and equal or ≥10 points indicates poorly-controlled and well-controlled disease, respectively, with 76% sensitivity and 84% specificity. 10Recently, the minimal clinically important difference (MCID) of AECT-4wk was determined to be three points; however, it has not been assessed for the AECT-3mth yet. 11e AECT is easy to perform and score, valid, and reliable, and it has become the gold standard for assessing disease control and guiding treatment decisions in RAE.It is widely available, in 47 languages including Turkish, but the Turkish AECT versions remain to be assessed for their validity, reliability, screening accuracy, sensitivity to change, and MCID.Therefore, we determined these clinicometric properties of the Turkish AECT-4wk and AECT-3mth versions and prospectively assessed them for their clinical utility across different types of RAE, that is, MMAE and BMAE including HAE.

| Study design
The study included 89 adult patients with RAE followed up at Istanbul Faculty of Medicine Adult Allergy and Immunology Clinic, Koc University Faculty of Medicine Dermatology Clinic and Sakarya University Faculty of Medicine Dermatology Clinic, all of which are members of Turkish Angioedema Center of Reference and Excellence centers. 12,13e study was conducted in two phases: linguistic and content validations.After the completion of the linguistic validation, both tools were administered to patients diagnosed with either MMAE or BMAE.

| Patient population
Patients' diagnosis was confirmed before enrollment according to international guidelines. 2,3All MMAE patients had acquired RAE, and all BMAE patients had HAE type 1 or type 2. Age and education level were similar between groups.The Female/male ratio was higher in MMAE group than in the HAE group (4.1 vs. 1.53, p = 0.039).The detailed demographic and clinical features of the patients are shown in Table S1.
[16] Meanwhile, all patients were assessed by physician expert on RAE and treated according to current international guidelines throughout the study period. 2,3e study was approved by the local ethics committee of Istanbul University and written informed consent was obtained from all participants.

| Translation and linguistic validation of the Turkish AECT versions
In order to develop a Turkish version of both AECT-4wk and AECT-3mth, the original German versions were translated into Turkish according to the structured translation process with a recommended road map consisting of the forward translation, the production of a "reconciled" version, backward translation, backward translation review, development of a consensus version and cognitive debriefing.
The pre-final Turkish versions which were developed according to consensus between the original and the Turkish authors were cognitively debriefed in four and three patients with MMAE and BMAE/ HAE, respectively.Four of them were females.The median (interquartile range [IQR]) age and elapsed time of interview were 44 (28-59) years and 12 (10-15) minutes, respectively.As a result of debriefing, one word was changed in question three to improve understandability.These final versions (Figure S1) were consented by the original developers and subsequently used for the content validation.

| Content validation of the Turkish AECT versions
The patients were assessed for their eligibility in the study at visit zero.If the patients accepted to be a participant, they were informed about the study, the demographic and clinical features were recorded and they were invited for further three visits (visit 1, visit 2 and visit 3).The intervals between visit 0 and 1, visit 1 and 2 (for the assessment of AECT-4wk) were 4 weeks and visit 1 and visit 3 (for the assessment of AECT-3mth) was 3 months.The final Turkish versions of AECT-4wk were applied to patients in visits 1 and 2 (4 weeks interval) and AECT-3mth was applied to participants in visits 1 and 3 (3 months interval).

| Assessment of disease activity, quality of life and disease control with other tools
AAS-28 day which was validated into Turkish, is a kind of angioedema diary that enables the patients self-record of angioedema attacks once daily. 6In visit 0 and visit 1, patients were asked to complete the angioedema activity score for 28 days (AAS-28 day) for the next visits to evaluate the disease activity.In addition to AAS-28 day, patients were asked for self-assessment of their disease control level for the last 4 weeks and 3 months with visual analog scaleangioedema control (VAS-AEC-4wk, VAS-AEC-3mth) and Likert scale-angioedema control for the last 4 weeks and 3 months (LS-AEC-4wk, LS-AEC-3mth).In VAS-AEC, patients were asked to mark a point on a continuous line ranging from 0 to 10 points.Zero points indicated that the disease was totally under control, while 10 points indicated poor disease control with maximum complaints. 15LS-AEC is a 5-point scale and indicates the global level of disease control comprising "completely controlled", "well-controlled", "moderately controlled" "hardly controlled" and "not at all controlled" (10).The treatment sufficiency from the perspective of patients was evaluated with a two-answer scale including "sufficient treatment" and "not sufficient treatment" for both the last 4 weeks and 3 months. 10her anchor tools including a two-choice scale for the assessment of the sufficiency of treatment from the patients' perspective and short form-12 (SF-12) and AE-QoL surveys for the determination of disease burden were applied to patients in visits 1-3.AE-QoL is a specific, valid and reliable survey for the assessment of the disease burden on patients with RAE. 6 It is composed of four domains and 17 questions and the total score ranges from 0 to 100, with a lower score indicating less impairment in QoL.Also, the impairment in the QoL related to RAE is classified as no relevant impairment (≤30.5),mildly impaired (30.6-37.9),moderately impaired (38-50.5)and severely impaired (≥50.6)according to total score. 17-12 is a nonspecific survey and is used in chronic diseases to indicate the impairment in the general health status.It is composed of physical component summary (PCS-12) and mental component summary (MCS-12) and the total score ranges from 0 to 100 points with higher scores showing better self-reported health status. 14

| Convergent validity
Convergent validity, that is, the extent to which responses on a test show a strong relationship with responses on conceptually similar tests, of the Turkish AECT-4wk and AECT-3mth were determined by Spearman's correlation analysis with the AAS-28 day, VAS-AEC, LS-AEC, AE-QoL and SF-12 components as anchor tools.Weak, moderate, and strong correlations were defined by correlation coefficients (r) of less than 0.3, between 0.3 and 0.5, and more than 0.5, respectively. 10

| Known groups validity
For the determination of known group validity, which shows the ability of the instrument to distinguish the patients with different disease states, we compared AECT versions with both LS-AEC, impairment level of QoL and the patients' self-assessment of treatment sufficiency by Kruskal-Wallis and Mann-Whitney U tests according to the type of the data. 10

| Test-retest reliability
Evaluation of stability and reproducibility of Turkish AECT versions was conducted in subgroups of patients with stable disease whose identical angioedema control levels according to LS-AEC-4wk and -3mth was detected in different visits.Interclass correlation coefficient (ICC) was calculated and classified as 0.5-0.7 indicates moderate to good reproducibility and more than 0.7 indicates excellent reproducibility. 18

| Screening accuracy
For the identification of the patients with poorly-or well-controlled disease, LS-AEC-4wk and -3mth were used as anchor instrument.
LS-AEC was further classified as well-controlled including the original answers of well and completely controlled and poorly-controlled including the original answers of not at all, hardly and moderately controlled.Receiver operating characteristic (ROC) analysis and the area under the curve (AUC) value were used to show the screening accuracy of the Turkish AECT versions.No discrimination, acceptable, excellent and outstanding screening accuracy was defined when the AUC values were 0.5, 0.7-0.8,0.8-0.9 and >0.9, respectively. 19ter that, patients were categorized as poorly-and well-controlled when AECT score ≤9 and ≥10, respectively.The median scores of AAS-28 day, number of attacks, AE-QoL and SF-12 components, impairment level of QoL and treatment sufficiency were compared between these two groups.

| Sensitivity to change
To determine the ability of Turkish AECT versions to detect changes in the patients' disease control, we performed Spearman's correlation analysis between the intra-individual changes of AECT scores and the other tools determined in two visits showing the changes in disease activity (AAS-28 day) and control level (VAS-AEC) and impairment in QoL (AE-QoL, PCS-12 and MCS-12) from visit 1-2 for AECT-4wk and from visit 1-3 for AECT-3mth.The correlation degree was categorized as defined in the convergent validity analysis.Moreover, a nonparametric t-test was used to show the relationship between the change in AECT scores and change in patient-self assessment of treatment sufficiency.

| Minimal clinically important difference
MCID indicating the meaningful changes in the patient's state over a period of time is an important feature of PROMs.We determined the MCID of Turkish AECT versions by an anchor-based method using LS-AEC. 20The patients were grouped into individuals with at least one step change in angioedema control level and with no change and also with at least one step of improvement (meaningful improvement) and with no improvement in angioedema control level according to LS-AEC.To identify the best cut−off point for MCID in AECT, ROC curve analysis was performed. 16The AUC value was categorized as told in the screening accuracy part.

| Statistical analysis
IBM SPSS statistics version 22 was used for the analysis of the data.
Descriptive data were given as number and percentages or median and IQR according to the type of the data.Categorical data were compared with Chi-square or Fisher's Exact tests depending on the distribution of the data.Continuous variables were compared with Kruskal-Wallis or Mann Whitney U test between groups according to the number of groups.Continuous data determined in different visits for the same scale were compared with a nonparametric paired t test.
Spearman's correlation test was used for the correlation analysis.
Statistical significance level was determined when p value was greater than 0.05.

| RESULTS
Fifty-one and 38 patients with MMAE and BMAE (all HAE) were enrolled, respectively.Eight patients did not attend visits 2 and 3.

| Convergent validity
The Turkish AECT-4wk was strongly correlated with AAS-28 day, VAS-AEC-4wk and AE-QoL scores and moderately correlated with the number of attacks in the last month and last 3 months, PCS-12, and MCS-12.The Turkish AECT-3mth was strongly correlated with AE-QoL and VAS-AEC-3mth and moderately correlated with AAS-28 day, number of attacks in the last 3 months, PCS-12, and MCS-12 (Table 1).These results demonstrate a good level of convergent validity for both Turkish AECT versions of AECT.

| Known-groups validity
The median AECT-4wk and -3mth scores were significantly different in patients with different levels of disease control and impairment of QoL according to the LS-AEC and AE-QoL survey (p < 0.001).Also, median AECT-4wk and AECT-3mth scores were higher in patients who reported that their treatment was sufficient (p < 0.001; p = 0.012) (Table 2).Taken together, these results demonstrated that both Turkish AECT versions were able to discriminate patients with various disease states and have high known-group validity.

| Internal consistency reliability
Both Turkish AECT versions showed excellent internal consistency reliability with a Cronbach's alpha value of 0.89 for the AECT-4wk and 0.83 for the AECT-3mth.In separate analyses of the AECT-4wk and -3mth in patients with MMAE or BMAE/HAE, Cronbach's alpha values were 0.87 and 0.81 and 0.9 and 0.86, respectively, indicating excellent disease type-specific internal consistency.

| Test-retest reliability
The Turkish AECT-4wk and -3mth were analyzed for test-retest reliability in 37 and 27 patients for whom disease control levels according to LS-AEC-4wk and -3mth were the same in two different visits with recall periods of 4-week and 3-months, respectively.The mean � SD AECT-4wk and AECT-3mth scores were similar in both visits (AECT-4wk: 10.57 � 4.65 vs. 10.76 � 4.66, p = 0.577; AECT-3mth: 8.37 � 4.1 vs. 9.15 � 3.68, p = 0.076).The ICC values, which were 0.88 and 0.84 in single measures and 0.94 and 0.91 in average T A B L E 1 Convergent validity of the Turkish AECT-4wk and AECT-3mth.

| Screening accuracy
The ability of the AECT versions for the detection of meaningful or nonmeaningful disease control was tested.In visit 1, 44 and 45 patients were categorized as having poorly-controlled and well-controlled disease according to LS-AEC-4wk and 60 and 29 patients had poorly-controlled and well-controlled disease according to LS-AEC-3mth, respectively.In the ROC curve analysis, the AUC value for the AECT-4wk was 0.853 and the AUC value for the AECT-3mth was 0.807, showing excellent screening accuracy for both Turkish AECT versions (Figure 1).The sensitivity and specificity of the Turkish AECT-4wk and-3mth to identify patients with poorly or well-controlled disease was best for ≥10 as the cut-off point (sensitivity:71.1%;specificity: 81.8%;Table 3).This cut-off was also best for the AECT-3mth (sensitivity: 85%; specificity: 65.5%)These results confirm that AECT scores equal or greater than 10 indicate well-controlled disease.
For the Turkish AECT-4wk, the median number of attacks in the previous month and the median scores of AAS-28 day, VAS-AEC-4wk, and AE-QoL were higher whereas the median MCS-12 and PCS-12 scores were lower in patients with poorly-controlled disease than with well-controlled disease (for number of attack p = 0.001, for PCS-12 p = 0.003, for others p < 0.001).Also, the number of patients whose QoL was moderately or severely affected according to AE-QoL was higher and the number of patients who reported sufficient treatment was lower in patients with poorly-controlled as compared to wellcontrolled disease (p < 0.001).For the AECT-3mth, the median number of attacks in the last 3 months and the median scores of VAS-AEC-3mth and AE-QoL were higher and the median score of PCS-12 was lower in patients with poorly controlled RAE as compared to wellcontrolled RAE (for number of attack p = 0.006, for VAS-AEC-3mth, AE-QoL and PCS-12 p < 0.001).Although the median MCS-12 score was numerically lower in patients with poorly-controlled disease, it was not statistically significantly different as compared to patients with well-controlled disease (p = 0.053).The number of patients who reported sufficient treatment was lower, whereas the number of patients with moderately or severely affected QoL was higher in patients with poorly-controlled disease than in those with well-controlled disease (p < 0.001, p = 0.003) (Table 4).Taken together, these results demonstrate that Turkish AECT scores equal or greater than 10 indicate well-controlled disease.

| Sensitivity to change
To evaluate whether the Turkish AECT versions can measure changes in the patients' disease status, we performed correlation

| Minimal clinically important difference
To determine the MCID of the Turkish AECT-4wk and AECT-3mth, we determined their average change in 32 patients who reported relevant improvement of disease control as assessed by the LS-AEC, over 4 weeks and 3 months.
The mean � SD and median (IQR) of change in AECT-4wk was 3.7 � 3.0 and 4 (2-5.75),respectively (Table S2).By ROC curve analysis, the cut-off points for meaningful AECT-4wks improvement with the best balance of sensitivity and specificity were two and three points (Figure 2A and For AECT-4wk and AECT-3mth the number of attacks was counted in the last month and in the last 3 months, respectively.b For AECT-4wk and AECT-3mth VAS-AEC-4wk and VAS-AEC-3mth were analyzed, respectively.
DEMIR ET AL.

| Comparison of AECT-4wk and -3mth
The AECT-4wk and -3mth scores were strongly correlated with each other in all patients with RAE as well as in patients with MMAE and HAE (Figure 3).
The median AECT-4wk was lower and VAS-AEC-4wk was higher in patients with HAE than in those with MMAE (p = 0.018, p = 0.001), whereas the other scores were similar (Table S1).

| DISCUSSION
The current study comprising patients with two major RAE types indicates that the Turkish AECT-4wk and-3mth versions are valid and reliable instruments for the assessment of control of RAE.
Importantly, our study is the first to evaluate the sensitivity to change and the MCID of the AECT-3mth and it provides novel insights on the clinicometric properties of the AECT in different RAE types.
The current study indicated that both the Turkish AECT-4wk and  Since AE-QoL is a disease specific measure of impairment in QoL while SF-12 is a nonspecific tool, this fact could explain the difference in the power of correlation between AECT and AE-QoL or SF-12 components.Similarly, in the original validation study, correlation with SF-12 was lower than the correlation with AE-QoL. 10In the Thai version, similar to our study, a good level of correlation with AE-QoL, AAS-28 day and numeric control scales was reported. 21Also, the Turkish AECT-4wk and -3mth had a high known-groups validity level, indicating that AECT can discriminate patients with various disease states in concordance with the original and Thai versions. 10,21The current study showed that the Turkish AECT versions were reliable tools with excellent internal consistency and reproducibility in line with the other versions. 10,21 observed a good level of screening accuracy in concordance with the original and Thai versions. 10,21In the original study, it was found that equal or greater than 10 points showed well-controlled disease with 75.5% and 81.8% sensitivity and 83.9% and 91.7% specificity for the AECT-4wk and -3mth, respectively. 10In the Thai study, the cut-off point to determine well-controlled disease was similarly equal or more than 10 points with 97.2% and 82.8% sensitivity and 88.9% and 77.8% specificity for the AECT-4wk and -3mth, respectively. 21We also observed that the best cut-off point seemed to be 10 with 71.1% and 85% sensitivity and 81.8% and 65.5% specificity for the AECT-4wk and -3mth, respectively.
Although the cut-off point was the same in all versions, our findings regarding sensitivity and specificity were more compatible with the original German versions than the Thai version, which could be explained by the geographical and ethnical variations.Then, we showed that the number of attacks, AAS-28 day (for AECT-4wk), VAS-AEC and AE-QoL scores were higher and PCS-12 and MCS-12 scores were lower in patients with poorly-controlled (AECT ≤ 9).Also, moderately or severely impairment in QoL and insufficient treatment were prominent in patients with poorly-controlled disease.
These findings indicated that AECT is concordant with other valid and reliable PROMs, supporting our result.Similar findings were also reported in the original study, while these analyses were not performed in the Thai study. 10,21Consequently, it is clear that an AECT score equal or less than 9 can detect poorly-controlled disease to aid the treatment decision.
In  -9 of11 respectively.In our study, we analyzed the MCID of AECT-3mth for the first time and showed that the 3-point increase in AECT-3mth can identify a meaningful improvement in disease control with 73.8% specificity.In HAE and MMAE, the specificity of 3-point increase in AECT-3mth was 87.5% and 55.6%, respectively.Therefore, we can say that the 3-point increase in AECT-4wk is a more favorable tool to determine the improvement in the disease control in all patients independent of the underlying diseases.On the other hand, 3point increase in AECT-3mth can be an indicator of meaningful improvement in control of HAE more specifically than in control of MMAE.
Regarding all the results, we can present that AECT-4wk can be a more accurate tool than the AECT-3mth to determine the control level of the disease even though we observed strong correlations between AECT-4wk and -3mth in all patients and also in patients with either HAE or MMAE.However, the other authors in Thai versions reported that AECT-4wk may be preferable to the AECT-3mth. 21 a limitation, we could not determine the impact of AECT on the deterioration of the disease control since we did not have enough patients with deterioration in disease control.Furthermore, we did not include other types of RAE such as idiopathic or acquired AE because they are rarely seen compared to HAE and MMAE and we anticipated that we could not obtain enough data during the study period to compare other groups.
To conclude, we found that the Turkish AECT versions were valid and reliable tools for the monitoring of disease control in patients with RAE.Equal or more than 10 points in AECT can indicate wellcontrolled disease.Three-point chance in AECT-4wk and -3mth scores can represent a minimal improvement in disease control level in patients with RAE.AECT-4wk seems to be more appropriate than the AECT-3mth for the evaluation of control and to detect the minimal clinically important change in RAE.We believe that the findings of our study will add to the utility of AECT in clinical practice.

-
3mth versions have a good level of convergent validity.Both versions showed strong correlation with AE-QoL and VAS-AEC scores and moderate correlation with SF-12 components and number of F I G U R E 2 ROC curve analysis to determine the relationship between at least one step improvement in disease control level and change in (a) AECT-4wk and (b) AECT-3mth.AUC, area under curve; ROC, receiver operating characteristic.T A B L E 5 Sensitivity and specificity for different cut-off values of the change in the Turkish AECT-4wk and -3mth versions leading to meaningful improvement, respectively.
the present study, we analyzed the sensitivity to change and MCID of AECT, as well.Very recently, Fijen et al. reported that the strength of sensitivity to change of AECT-4wk to detect the change in disease control determined by physicians, angioedema related QoL and patient rated treatment sufficiency was high while it was lower in the detection of change in patients-self assessment of disease control.However, they did not analyze the correlation between change in AECT-4wk and change in disease activity and sensitivity to change in AECT-3mth. 11We found that changes in AECT-4wk were sensitive to detect the changes in disease activity, disease control assessed by patients and treatment sufficiency, while less sensitive to detect the changes in AE-related QoL.The sensitivity to change of AECT-3mth to detect the change in disease control and AE-Qol was weak.Furthermore, correlations between changes in AECT versions and changes in SF-12 components were not substantial.Fijen et al. also reported a low level of correlation between the change in AECT-4wk and dermatology life quality index (DLQI).
Demir, Deniz Eyice Karabacak, Derya Ünal, İlkim Deniz Toprak, Pelin Korkmaz, Ayşe Feyza Arslan, Işıl Göğem İmren, Bahar Dikicier, Nevzat Kahveci, Nida Öztop, Rabia Öztaş Kara and Aslı Gelincik made substantial contributions to conception and design, acquisition of data, and analysis and interpretation of data; and were involved in drafting the manuscript or revising it critically for important intellectual content; and gave final approval of the version to be published.Emek Kocatürk, Marcus Maurer and Karsten Weller made substantial contributions to conception and design and interpretation of data, were involved in revising the manuscript critically for important intellectual content, and gave final approval of the version to be published.Each author participated sufficiently in the work to take public responsibility for appropriate portions of the content; and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Table 5
Sensitivity and specificity for various cut-off points of the Turkish AECT-4wk and -3mth to discriminate poorly-and well-controlled disease in patients with recurrent angioedema.Comparison of number of disease activity, impairment in QoL and treatment sufficiency between patients with poorly-and well-controlled disease.
).When we performed the same analysis in MMAE (n = 22) and BMAE/HAE (n = 17) separately, a T A B L E 3 a