Article Text
PDF
Abstract
Objective Glucocorticoids (GC) are a cornerstone in treating antineutrophil cytoplasmic antibodies-associated vasculitides (AAV), however, they add to morbidity and mortality. To date, GC toxicity in AAV has rarely been systematically investigated.
Methods Patients with a confirmed AAV were included in this monocentric prospective study. GC toxicity was assessed by structured interviews, clinical examination and electronic medical record analysis. The Glucocorticoid Toxicity Index (GTI) consisting of the Aggregate Improvement Score (GTI-AIS) and the Cumulative Worsening Score (GTI-CWS) was assessed at two time points (t1 baseline, t2 6 months later). We used regression analyses to assess the relationship between GTI and GC exposure, toxicity, and disease activity, and a receiver operating characteristic analysis to calculate a GC threshold dose beyond which toxicity is expected to occur.
Results We included 138 patients with AAV. The median cumulative GC dose was 9014.0 mg. The most frequent adverse events were skin atrophy, osteoporosis and myopathy. GC exposure and toxicity were significantly correlated (p<0.001). GTI-AIS was significantly higher in active disease compared with patients in remission (p<0.001). GTI-CWS scored significantly higher in long-standing diseases (p=0.013) with high cumulative GC doses (p=0.003). Patients with a cumulative GC dose of 935 mg or more showed an 80% likelihood for a clinically meaningful change in GTI scoring.
Conclusion The GTI is capable of capturing GC toxicity in AAV and identifies patients at increased risk for GC side effects. Our data support efforts to limit GC exposure in patients with AAV.
- Glucocorticoids
- Granulomatosis with polyangiitis
- Systemic vasculitis
- Vasculitis
Data availability statement
Data are available on reasonable request. On reasonable request to the corresponding author, source data for this work can be made available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
The Glucocorticoid Toxicity Index (GTI) has been validated in asthma patients. Its value in antineutrophil cytoplasmic antibodies-associated vasculitis (AAV) remains to be evaluated.
WHAT THIS STUDY ADDS
We provide prospective data in 138 AAV patients evaluating the GTI. The data underline the commonness and severity of glucocorticoids (GC) toxicity and the usefulness of the GTI in capturing it. The threshold GC dose beyond which toxicity becomes likely was calculated as 935 mg prednisolone. The GTI scores positively correlate with the domain ‘physical limitation’, but not the ‘treatment side effects’ of the AAV-specific patient-reported outcome questionnaire.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Since GC toxicity is high and frequent in AAV patients, novel therapies should aim at reducing their use. In order to achieve this goal, a reliable instrument to measure and objectively quantify GC toxicity in AAV is necessary. Therapy trials should include the GTI as an endpoint measurement tool.
Introduction
Antineutrophil cytoplasmic antibodies (ANCA)-associated vasculitides (AAV) encompass granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA) and eosinophilic granulomatosis with polyangiitis (EGPA) according to the revised nomenclature from the 2012 Chapel Hill Consensus Conference.1 2 Their hallmark is a necrotising inflammation preferentially affecting small vessels. ANCA specifically directed against proteinase-3 or myeloperoxidase are often detectable, and immune complex deposits are rare.1 The annual incidence in Germany is 34 cases per 1 million for GPA, 13 cases per 1 million for MPA and 2 cases per 1 million the EGPA.3–5
Despite modern immunosuppressants such as rituximab, mepolizumab, avacopan and others in both induction and maintenance therapy, glucocorticoids (GC) still play a crucial role in the treatment of AAV and are part of current guideline recommendations.6–8 GC substantially add to the significantly higher mortality of AAV patients compared with the general population.9–12 Severe courses of disease and frequent relapses lead to high cumulative GC doses with a considerable risk for GC-associated side effects.9
GC-associated complications and sequelae in patients with AAV have rarely been systematically studied.13 14 Before recently, GC side effects have not been assessable in a standardised fashion. The Glucocorticoid Toxicity Index (GTI) is a novel clinical outcome measurement assessing the impact of GC therapy over time and its change between two points in time. It was validated in patients with asthma and has recently been used in a prospective clinical trial in AAV.12 In our study, we prospectively investigated the impact of GC toxicity using the GTI in routine care in a large AAV cohort at two points in time.
Patients and methods
Study design and subjects
In this prospective monocentric study, subjects with newly diagnosed and prevalent AAV were recruited from our department, which is a tertiary referral centre for vasculitis.
Baseline data were collected for each subject by reviewing electronic medical records and interviewing and examining subjects in person using a standardised form. We recorded AAV subtype, stage and activity state, organ manifestations, ANCA subtype, disease duration, treatment, GC dose and AAV relapses.
Disease activity of AAV was assessed using the Birmingham Vasculitis Activity Score version 3 (BVAS v3),13 and disease-related and therapy-related damage was measured by the Vasculitis Damage Index in patients with AAV.15
We collected data on two time points, t1 represented baseline, t2 6 months later.
We defined the ‘total cumulative GC dose’ as the sum of applied GC doses (prednisone equivalent) over the entirety of the disease. It was calculated by analysing electronic medical records over the entire disease duration. For any data before t1, acquisition of GC dose was retrospective. Standard of care protocols and medical records were used to determine GC exposure throughout the entire course of disease.
We further defined the ‘current GC exposure’ as the sum of the applied GC doses during the prospective observational period between t1 and t2.
GC toxicity measurement
GC-related toxicity was assessed using the GTI 2.0, a validated and standardised tool to capture the change in GC toxicity over time. The GTI consists of 9 domains and 47 items. Each item has an individual weighting according to the toxicity depicted. In all domains, there is an option to score improvement, worsening or no change.16 The GTI Aggregate Improvement Score (GTI-AIS) captures current GC toxicity and can reach values from −346 to 439. Negative values show improvement and positive values show worsening of GC toxicity. In our study, the GTI-AIS indicated the change in GC toxicity during the 6-month period between t1 and t2, t1 reflecting the GTI-AIS baseline and t2 being 6 months later. The GTI Cumulative Worsening Score (GTI-CWS) captures cumulative GC toxicity and can only increase or remain on its current value over time. Therefore, the GTI-CWS is always ≥0, its maximum value is 439. The GTI-CWS was calculated for the entire disease duration ranging between the time of the initial diagnosis and t2.16 According to literature, we defined the minimal clinically important difference (MCID) in change of GC toxicity as a GTI≥10.17
GC-related complications between t1 and t2 were assessed by analysing electronic medical records and interviewing patients. In addition, each patient underwent a physical examination. GC toxicity items were based on a defined list of GC-related side effects (online supplemental table S1).11
Supplemental material
To capture the patients’ perspective regarding their disease, we used the validated AAV patient reported outcomes (AAV-PRO) questionnaire17 18 and correlated questionnaire scores with GTI results.
Statistical analyses
Data were analysed by using SPSS V.28 (IBM). Categorical variables were expressed in counts and frequencies (n (%)). Regarding their distribution, continuous variables were expressed as means±SD if distributions were relatively normal or as medians with IQR in cases of skewed distribution. The values represent the conditions at t2.
For continuous variables, the significance of differences across subgroups was analysed according to their distribution using t tests or Mann-Whitney U tests. To assess the significance of association between categorical variables and subgroups, the χ2 statistic was calculated. In the case of low predicted cell counts for categorical variables, Fisher’s exact test was applied.
To visualise the associations between GC exposure and GTI scores, we constructed a scatter plot and calculated a linear regression model. In addition, Spearman’s r correlation coefficient was calculated.
To calculate a threshold value beyond which GC doses will cause any change and a clinically meaningful change in the GTI (MCID), receiver operating characteristic (ROC) analyses were performed and the Youden index was assessed. For the MCID, a GTI-AIS>10 (a suggested in Stone et al19) was used as the state variable while the current GC exposure between t1 and t2 served as the test variable.
AAV patients with active disease had a BVAS>0 at t1, whereas AAV patients in stable remission were defined by a BVAS of 0.
Results
Patient characteristics
For this study, 138 patients with AAV were recruited (table 1). AAV patients included 71 patients with GPA (51.4%), 25 patients with MPA (18.1%) and 42 patients with EGPA (30.4%). The mean age was 56.8 (±15.3) years.
GC toxicity according to disease activity
The majority of patients were women (n=76, 55.1). The median current GC exposure was 501.5 mg (IQR 0.0–993.3). The mean total cumulative dose was 9014.0 mg (IQR 4974.8–17 389.3). The median disease duration was 57.0 months (IQR 25.8–103.5).
29 patients (21.0%) had active disease (BVAS>0) at t1, while 109 patients (79.0%) were in stable remission during the same period (BVAS=0).
During the observational period between t1 and t2, an infection was recorded in 10% of patients. Over the entire disease duration, 46 (33%) patients developed infections, 13.8% of which were severe.
Change in GC toxicity between t1 and t2
At the end of the observation period (ie, t2), 50.7% of our AAV patients showed no change in recent GC toxicity (GTI-AIS=0), in 27 (19.6%) an improvement (GIT-AIS<0) and in 41 (29.7%) an increase in GC toxicity (GTI-AIS>0) was documented. The current GC exposure between t1 and t2 correlated positively with GTI-AIS scores (r=0.490; p<0.001) as shown in figure 1, which was supported by regression analyses (adjusted R2=0.262; F(df=1;136)=49.74; p<0.001).
Cumulative GC dose and GTI-AIS scoring. Scatter plot showing the relationship between cumulative GC dose in mg and GTI-AIS scoring. AIS, Aggregate Improvement Score; GC, glucocorticoids; GTI, Glucocorticoid Toxicity Index.
Of those patients whose GC toxicity worsened during t1 and t2 19 (46%) had active AAV. In patients with active disease, 65.5% had an increasing GC toxicity (GTI-AIS>0) compared with 20.2% in those with stable remission (p<0.001). In parallel, the mean current GC exposure in active disease versus remission was 1745.0 mg (IQR 1206.5–2542.0) vs 240.0 mg (IQR 0.00–733.00), respectively (p<0.001). A complete list of features characterising patients with active disease and remission is shown in table 1.
Overall GC toxicity
The median GTI-CWS at t2 was 58.5 (IQR 29.0–102.3). 11 patients (8.0%) had a GTI-CWS of 0, whereas in 127 (92.0%) the GTI-CWS was>0. In both correlation and regression analyses, the GTI-CWS showed a weak association with the total cumulative GC dose (Spearman’s r=0.251; p=0.003. Regression: adjusted R2=0.262; F(df=1;136)=49.74; p<0.001).
We found a weak relationship between disease duration and GTI-CWS (Spearman’s r=0.215, p=0.11).
Neither the current GC exposure between t1 and t2 (r=0.042; p=0.556) nor the disease activity state at t1 (p=0.664) exerted any influence on the GTI-CWS (figure 2).
Percentage of patients with GTI-AIS 0, >0 and <0 according to disease activity. In patients with active disease changes in current GC toxicity as captured by the AIS (GTI-AIS) are more likely compared with patients in stable remission. AIS, Aggregate Improvement Score; GC, glucocorticoids; GTI, Glucocorticoid Toxicity Index.
Sex-related and age-related differences in GC toxicity
The total cumulative GC dose and the current GC exposure between t1 and t2 showed no significant difference between women and men, nor did the acute change in GC toxicity (online supplemental table S4). Cumulative GC toxicity (CWS) was significantly higher in women than in men. Osteoporosis was found more frequently in women (p=0.022) while acne was more common in men (p=0.022).
Patients of 65 years and above were more prone to cataract (p=0.035), skin atrophy (p=0.005) and diabetes (p=0.029). All other items of GC toxicity showed no differences when dividing the cohort by sex and age (online supplemental tables S3 and S4).
GC threshold dose for GC-related toxicity
The ROC analysis with the current GC exposure between t1 and t2 and the GTI-AIS (figure 3) showed an area under the curve (AUC) of 0.815 (95% CI 0.733 to 0.896) with a threshold dose of 733.0 mg for the state variable GTI-AIS>0 and a threshold dose of 935 mg considering the MCID (GTI-AIS>10) with an AUC of 0.800 (95% CI 0.713 to 0.877).
Determination of GC threshold dose. Receiver operating characteristics (ROC) curves to determine the cumulative GC threshold dose between t1 and t2 beyond which any GTI-AIS change (A) and a GIT-AIS>10 (B) representing the minimum clinically important difference. (A) AUC 0.815, p<0.001, 95% CI 0.733 to 0.896. (B) AUC 0.800, p<0.001, 95% CI 0.713 to 0.887. AIS, Aggregate Improvement Score; AUC, area under the curve; GC, glucocorticoids; GTI, Glucocorticoid Toxicity Index.
Association of GC toxicity and PROs (AAV-PRO)
AAV patients with GTI-AIS>10 had a significantly higher score in the AAV-PRO domain ‘physical functioning’ compared with AAV patients with GTI-AIS<10 (p=0.034). Patients with active disease scored significantly higher in the ‘concerns about the future’ category compared with AAV patients in stable remission (p=0.019). All other dimensions of the AAV-PRO did not significantly correlate with GTI results, including the domain ‘treatment side effects’. The complete data on AAV-PRO and GTI-AIS are shown in table 2.
Comparison between GTI and AAV-PRO questionnaire scoring
Discussion
This prospective study investigated GC-associated complications and sequelae in a real-world cohort of patients with AAV using the GTI as a validated scoring instrument for GC toxicity.
Our data confirm that GC-associated complications in AAV patients are very common, reflected by only 8% of patients with a GTI-CWS of 0. AAV disease duration showed only a very weak correlation with GTI-CWS scoring. In our cohort, patients with stable remission had significantly longer disease durations compared with those with active AAV, which might have affected these findings.
In this study, we found relatively low infection rates in our cohort, especially between t1 and t2. Published data report that 30% of AAV patients develop severe infections requiring intravenous antimicrobial treatment and/or hospitalisation.9 20 In general, GC treatment is associated with increased rates of infections,9–11 which led to several endeavours to reduce GC exposure in AAV patients. However, results of studies investigating GC-sparing strategies with regard to the risk of infections are heterogeneous: In the PEXIVAS trial, a reduced GC protocol was capable of significantly diminishing severe infectious events by 31%.21 In the LOVAS trial, a reduced GC protocol was capable of reducing infections by almost 13%.22 In contrast, reduction of GC exposure in patients receiving avacopan in a phase 3 trial was associated with an overall reduction of GC toxicity after 26 weeks but did not lead to significantly reduced infection rates.12 In a study with EGPA patients, mepolizumab led to a drastic reduction in GC use, but GC toxicity measured by the GTI-AIS did not change significantly in the overall cohort.23 Prophylaxis with trimethoprim/sulfamethoxazole was applied in 29 (21%) of our AAV patients. In patients with a mean daily prednisone dose of >7.5 mg 50% received prophylaxis, compared with 16% of those with a prednisone dose of ≤7.5 mg. This might have added to the low infection rates observed in our cohort.24
It was to be expected that GC toxicity is dose-dependent, which our data confirm. However, correlations between GC dose and GTI-CWS were surprisingly weak. EGPA patients showing the highest cumulative GC dose in our study did not reach higher toxicity scores in the GTI-CWS. However, GTI-AIS was significantly influenced by disease activity, which is consistent with dose-dependent GC toxicity since patients with new onset disease or flares usually require higher GC doses.
Interestingly, 20% of our patients had an increase in recent GC toxicity during the observational period between t1 and t2 (GTI-AIS>0) despite being in complete remission and on a low GC dose (mean cumulative dose of 240 mg prednisone between t1 and t2). This supports our observation that even low daily GC doses cause toxicity when applied over a longer period of time. Furthermore, this might raise suspicion that other factors leading to health states captured by the GTI, especially cardiovascular diseases, are present.
We conclude that the relationship between GC use and objective measures of toxicity in AAV is complex, especially with regard to the MCID. Many GC toxicity items, especially the ones associated with cardiovascular diseases, are affected by factors beyond vasculitis such as behavioural habits, socioeconomic factors, education and genetics. For this reason, diseases such as atherosclerosis, myocardial infarction, stroke or heart failure are not captured by the GTI. Still, active AAV and its therapy are likely to negatively affect these conditions. According to our findings, cumulative GC doses and disease activity significantly impact GC toxicity.
AAV patients with active disease who received higher GC doses for induction of remission were more likely to exhibit acute GC side effects such as weight gain, diabetes, sleep disturbances and arterial hypertension than AAV patients in stable remission. In contrast, long-term GC toxicities such as skin atrophy were significantly more common in patients in stable remission due to the high cumulative GC exposure over the entire treatment period.
Using an ROC analysis, we calculated that at a cumulative GC dose of 935 mg within 180 days is highly predictive of a clinically meaningful GC toxicity event, which to our knowledge, has never been reported thus far. Thus, an average dose of just over 5 mg prednisone per day applied over 6 months is sufficient to reach this threshold. Provided that a patient with life and organ-threatening AAV requires a GC pulse therapy with 250 mg prednisolone intravenously over 3 days followed by a GC taper regimen according to PEXIVAS protocol,21 the toxicity threshold derived from our data would be reached within 1 week of therapy. This is underlined by the observation that over 90% of AAV patients in our cohort exhibit GC toxicity in the GTI-CWS. Of note, this threshold certainly does not apply for all GC toxicities. Osteoporosis and skin atrophy, for instance, will only occur after a long-term GC exposure.
The data of Flossmann et al point out that older adults show a higher vulnerability to GC side effects in multiple organ systems,9 which is supported by our findings, especially regarding diabetes and skin atrophy.
It is noteworthy, that the novel AAV-PRO questionnaire showed some significant correlations with the GTI results. In our cohort, physical limitations were reported significantly higher in patients with high GC toxicity, whereas, surprisingly, no significant findings were observed in the domain ‘treatment side effects’. The discrepancy between patient and physician reported toxicity and damage is well known.25 26 Other authors describe similar findings using the European Quality of Life 5 Dimensions 5 Levels (EQ5D5L) instrument.23 This might be explained by the fact that the AAV-PRO purely reports PROs, while the GTI is a clinician-facing instrument relying on patient input.19 For instance, elevated blood pressure or hyperlipidemia can be detected by the physician but could easily be missed by a PRO instrument. Disease-specific PRO instruments have become a field of interest in rheumatology. However, our data suggest, that GC-related toxicity can easily be missed when relying solely on PRO, and therefore, argue for the use of the GTI in AAV patients.
Our study has limitations worth mentioning: It is a single-centre study and data were taken from routine care. The selection of control patients might have been biased. The (in part) retrospective assessment of the cumulative GC dose from medical records might be erroneous. Only 21% of our patients had active disease, which limits the interpretability for this subset of patients and might also explain the low rate of infections detected in our study. We only included established or patient-reported diagnoses that accounted for (possible) GC-related toxicities. There was no active screening for these conditions. Some correlations between GTI items and cumulative GC doses are surprisingly weak or non-existent, which might be due to the fact that we did not adjust for preexisting toxicities before the observational period.
In summary, AAV patients have a high exposure to GC over the course of their disease. Our findings show that nearly all AAV patients experience GC-associated toxicity. Our data support the urgent need to limit GC exposure in AAV patients, which can be achieved by developing and establishing efficacious immunosuppressive agents allowing the restriction of dose and time of GC exposure.
Data availability statement
Data are available on reasonable request. On reasonable request to the corresponding author, source data for this work can be made available.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and we obtained approval for the conduction of this study from the Ethics Committee of the Eberhard Karls University, Tübingen, Germany (project no. 740/2021BO2). Participants gave informed consent to participate in the study before taking part.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors PJS: project planning, data acquisition, drafting manuscript, statistical analyses. BH: project planning, drafting and revising manuscript. Y-SF: statistical analyses, drafting and revising manuscript. CL: data acquisition, drafting and revising manuscript, statistical analyses, guarantor.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.