Specific immunotherapy for allergic rhinitis to grass and tree pollens in daily medical practice—symptom load with sublingual immunotherapy compared to subcutaneous immunotherapy

Abstract

Background. Despite strong evidence for subcutaneous and sublingual immunotherapy for the treatment of allergic rhinoconjunctivitis, comparative data are scarce.

Objectives. We performed an individual patient data meta-analysis of four observational studies to compare the effectiveness of both application routes.

Methods. After individual analysis, a subsequent analysis of the total data pool was performed. Descriptive and explorative data analysis methods were used.

Results. Altogether 847 patients (382 male, 453 female) aged 3–78 years (mean age 28.3 years) were treated with specific immunotherapy: 665 (78.5%) patients sublingual and 182 (21.5%) subcutaneous. The majority of patients (61.6%) in both treatment groups started specific immunotherapy due to severe rhinitis symptoms which occurred frequently or very frequently. Most patients in both treatment groups had moderate to severe conjunctivitis symptom load which occurred frequently or very frequently. Median rhinitis and conjunctivitis symptom loads decreased during both treatments to the same extent. Similar improvements in the symptom loads were observed in patients stratified for age, disease duration, and presence or absence of mild to moderate asthma.

Conclusion. The effectiveness of sublingual and subcutaneous immunotherapy with pollen extracts appeared virtually equal in daily medical routine. Due to the advantageous safety profile, the sublingual application may be favorable.

Key words::

• Allergic rhinitis
• grasses
• subcutaneous immunotherapy
• sublingual immunotherapy
• trees

Abbreviations
ANOVA	=	analysis of variance
ARIA	=	Allergic Rhinitis and Its Impact on Asthma workshop
IPD	=	individual patient data
IR	=	index of reactivity
OD	=	once daily
RCT	=	randomized controlled trial
SIT	=	specific immunotherapy
SCIT	=	subcutaneous immunotherapy
SLIT	=	sublingual immunotherapy
SMD	=	standardized mean difference
SmPC	=	summary of product characteristics
WHO	=	World Health Organization

Key messages

• Efficacy and safety of subcutaneous or sublingual allergen-specific immunotherapy has been proven in numerous clinical trials.

• This individual patient data meta-analysis of prospective, observational trials demonstrates that the effectiveness of sublingual and subcutaneous allergen-specific immunotherapy appears to be comparable in daily medical routine.

• Due to the advantageous safety profile, the sublingual application may be favorable.

Introduction

The recommendation to use sublingual immunotherapy (SLIT) in children and adults with allergic rhinitis evolved over several years and is based on the following key publications: the World Health Organization (WHO) position paper on allergen immunotherapy (1), the Allergic Rhinitis and its Impact on Asthma (ARIA) Workshop Group paper in collaboration with the WHO (2) including its update (3), the Cochrane review of SLIT for allergic rhinitis (4), a meta-analysis evaluating the efficacy of SLIT in the treatment of allergic rhinitis in children and adolescents aged 3–18 years (5), and the position paper of the World Allergy Organization (6). There is no doubt that subcutaneous immunotherapy (SCIT) is also effective in children and adults (1,3,7), but it is burdened by the risks of side-effects that can be life-threatening (2).

Over the past 10 years, the allergen doses used for SLIT increased up to 1,125 times when compared with the dose commonly administered in SCIT (8). A meta-analysis showed that this high-dosed SLIT is safe (9).

There are randomized trials which indicate that SCIT and SLIT are comparable with regard to efficacy (10,11), one even with double-dummy technique which represents the gold standard for the comparison of different application routes (10), but the robustness of the data is limited due to the small sample sizes. Furthermore, controlled clinical trials that focus on efficacy and safety as their specific aim are often conducted under tightly controlled research conditions. These studies often involve very selected patient populations, the results of which cannot be extended to all patients with allergic rhinitis. Studies of clinical effectiveness have evolved in response to the need for more real-world information about treatment alternatives and patient outcomes.

Therefore, we performed an individual patient data (IPD) meta-analysis of three open, prospective observational studies on SLIT and one open, prospective observational study on SCIT using standardized allergen extracts (Staloral 300^® and ALUSTAL^®; Stallergenes, Antony, France) in patients with allergic rhinitis, to compare the effectiveness of SLIT and SCIT in daily medical practice.

Methods

Patient selection

All three SLIT studies and the SCIT study recruited patients with a medical history of allergic rhinitis or rhinoconjunctivitis and with or without mild to moderate asthma due to grass pollens (cocksfoot, meadow grass, rye grass, sweet vernal grass, timothy) or tree pollens (birch, alder, hazel). The patient selection was based on the recommendations in the summary of product characteristics (SmPC) of the allergen extracts. Positive skin prick tests to at least one of the aforementioned allergens and presence of specific IgE ≥ 0.7 kU/L were required.

Study design

Criteria for the inclusion of patients and the scores for the measurement of the outcomes were identical in all studies. The duration of the observational phase was one year in studies 1 to 3 and three years in study 4, with outcome measurements after each year.

Study 1: 233 patients aged 2–69 years were treated in Germany. SLIT titration started pre-seasonally at 0.1 IR (index of reactivity), reached the daily maintenance dose of 300 IR on day 12, and was followed by a co-seasonal maintenance phase at 240 IR once daily (OD) for 5 days a week (Monday to Friday).

Study 2: 224 patients aged 10–72 years were treated in Germany. Ultra-rush titration reached the maintenance dose within 90 min (30–90–150–300 IR) and was followed by a co-seasonal maintenance phase at 240 IR OD.

Study 3: 208 patients aged 3–78 years were treated in Germany, Switzerland, and Italy. SLIT titration started pre-seasonally at 1 IR, reached the daily maintenance dose of 240 IR OD on day 12, and was followed by a perennial maintenance phase.

Study 4: 182 patients aged 4–66 years were treated in Germany. SCIT titration started pre-seasonally at 0.001 IR, reached the monthly maintenance dose of 8 IR after 13–17 weeks, and was followed by a perennial maintenance phase.

Patients were asked for a retrospective assessment of their symptoms during the previous year. The post-treatment evaluation was limited to the first year of treatment in all studies.

At the end of this observation period, symptoms during specific immunotherapy (SIT) treatment were assessed. Rhinitis and conjunctivitis symptom scores consisted of six symptoms (sneezing, rhinorrhea, nasal pruritus, nasal congestion, ocular pruritus, and watery eyes). The asthma symptom score consisted of dyspnea and wheezing. No symptoms and mild, moderate, and severe intensity of the individual symptoms were scored as 0, 1, 2, and 3, respectively. The frequency of symptoms was described as rare, sometimes, frequent, and very frequent and scored as 1, 2, 3, and 4, respectively.

Study drug

Standardized allergen extract in aqueous solution was used for SLIT (Staloral 300^®) and standardized aluminum hydroxide-adsorbed allergen extract for SCIT (ALUSTAL^®). Biological activity of the extracts was assessed in comparison with an internal standard, in vitro and in vivo, and expressed as index of reactivity (IR) (12). Allergen extracts used contained allergens of the following groups: grasses (predominantly sweet vernal grass, cocksfoot, rye grass, timothy, meadow grass), and trees (birch, alder, hazel).

Statistical analysis

The data have been analyzed in an analogous manner as described for SLIT (13) and antihistamines (14). To avoid any bias upon data consolidating, only variables which were assessed in all four studies were chosen for the analysis file, i.e. rhinitis and conjunctivitis symptom scores during the year before SIT and during SIT, respectively. A predefined protocol described the processing of the different variable groups both during the synthesis process and subsequent to it. Following data synthesis, the data sets were examined with respect to plausibility and occurrence of errors. If it was not possible to rectify the variables of a data set, missing values were assumed. After plausibility control, patient data sets with less than 50% of all data necessary for the calculation of analysis parameters were planned to be excluded from the final analysis file. None of the data sets was actually excluded.

The following variables were available for analysis: 1) Demography (age, gender, duration of rhinitis, sensitization); 2) Allergen extract used for SIT (kind of allergens, start and end of titration phase, kind of titration); and 3) Symptom scores (severity and frequency of rhinitis and conjunctivitis).

The four studies were first analyzed individually. A subsequent analysis of the total data pool was used for descriptive statistics. In both cases, the statistical analysis was primarily based upon descriptive and explorative data analysis methods. The demographic data were checked for homogeneity among studies, and ordinal parameters were examined by means of contingency tables. Both Kruskal–Wallis test and Mann–Whitney U test were used to examine differences between studies.

During stratified analysis, the studies were used both individually as well as summarized in a total data pool. The evaluation of each study with a subsequent synthesis of the results served to estimate the overall effect. In the analysis of the total data pool, calculations of the overall effect and analysis of variance (ANOVA) for possible influencing factors were performed.

The total data pool was stratified as follows: 1) SLIT versus SCIT; 2) Patients with asthma versus patients without asthma; 3) Treatment with tree pollen extract versus treatment with grass pollen extract; 4) SLIT versus SCIT in the two subgroups of patients without and with asthma; 5) Various age groups (< 12; 12 to < 18; 18 to < 25; 25 to < 40; 40 to < 65; and ≥ 65 years of age); 6) Various durations of disease (≤ 1; 1 to < 2; 2 to < 4; 4 to < 5.76; 5.76 to < 11; and ≥ 11 years); and 7) SLIT versus SCIT in the two subgroups of patients treated with tree pollen extract versus patients treated with grass pollen extract.

The efficacy of individual strata was compared using two-sided statistical tests. Differences were considered statistically significant at a P value of less than 0.05.

Estimation of the study-specific effects and examination for homogeneity

The estimations of effects were based on symptom scores. These parameters are ordinal numbers. Nevertheless, procedures for continuous variables, a widely used approach for meta-analyses, were applied. Weighted mean differences were calculated and checked for homogeneity. Variability in patient characteristics and conduct of the studies were used to explain heterogeneous study-specific effects.

Estimation of the pooled effect

The overall effect was estimated using the ‘inverse variance’ method. For homogeneous and heterogeneous study-specific effects, the ‘fixed effect’ and ‘random effects’ models were used.

Pooled analysis

Significant differences in efficacy between strata appeared and were analyzed by ANOVA for possible influencing factors. A model with constant and random variable effects was applied. The variable ‘study affiliation’ constituted the random variable effect.

Results

A total of 847 patients (382 male and 453 female patients, 12 patients without gender information) aged 3–78 years (mean age 28.3 years) were treated with SIT: 665 (78.5%) patients with SLIT and 182 (21.5%) with SCIT. All patients suffered from rhinitis, and 291 had asthma (38.0% of 766 patients had information about asthma). Nearly half of the patients (n = 408; 48.2%) received grass pollen extract and 439 (51.8%) tree pollen extract. The mean duration of the disease was 6.3 years (range 0–46 years). More than half of the patients (n = 437; 51.8%) were monosensitized, and 406 (48.2%) were polysensitized (no information: n = 4). Continuation of treatment after the observation period was confirmed for 80%–90% of patients in the SLIT studies; 81% of patients continued the treatment in the second year in the SCIT study.

The severity and frequency of rhinitis and conjunctivitis symptoms before start of SIT (pre-treatment) and after 1 year of treatment (post-treatment) are summarized for SLIT and SCIT in Table I and Table II, respectively. The vast majority of patients in both treatment groups started SIT with severe rhinitis symptoms, which occurred frequently or very frequently (Table I). Most patients in both treatment groups started SIT with moderate to severe conjunctivitis symptoms, which occurred frequently or very frequently (Table II).

Table I. Severity and frequency of rhinitis symptoms.

		Severity of rhinitis symptoms
		None		Mild		Moderate		Severe
		Pts.	%	Pts.	%	Pts.	%	Pts.	%
Pre-treatment	SLIT	4	0.6	25	3.8	251	38.0	381	57.6
	SCIT	0	0	0	0	44	24.2	138	75.8
Post-treatment	SLIT	116	18.8	267	43.2	196	31.7	39	6.3
	SCIT	8	5.8	48	34.5	66	47.5	17	12.2

		Frequency of rhinitis symptoms
		Rare		Sometimes		Frequent		Very frequent
		Pts.	%	Pts.	%	Pts.	%	Pts.	%
Pre-treatment	SLIT	8	1.3	87	14.1	346	56.3	174	28.3
	SCIT	1	0.6	17	9.7	116	65.9	42	23.9
Post-treatment	SLIT	126	28.3	203	45.6	92	20.7	24	5.4
	SCIT	27	20.9	69	53.5	28	21.7	5	3.9

Table II. Severity and frequency of conjunctivitis symptoms.

		Severity of conjunctivitis symptoms
		None		Mild		Moderate		Severe
		Pts.	%	Pts.	%	Pts.	%	Pts.	%
Pre-treatment	SLIT	39	6.0	105	16.1	293	44.9	215	33.0
	SCIT	8	4.4	19	10.5	67	37.0	87	48.1
Post-treatment	SLIT	216	35.4	241	39.4	126	20.6	28	4.6
	SCIT	40	28.8	45	32.4	42	30.2	12	8.6

		Frequency of conjunctivitis symptoms
		Rare		Sometimes		Frequent		Very frequent
		Pts.	%	Pts.	%	Pts.	%	Pts.	%
Pre-treatment	SLIT	48	8.4	153	26.8	273	47.8	97	17.0
	SCIT	12	7.1	39	22.9	93	54.7	26	15.3
Post-treatment	SLIT	138	38.7	134	37.5	66	18.5	19	5.3
	SCIT	24	25.0	51	53.1	16	16.7	5	5.2

The median (range) rhinitis scores for severity and frequency of symptoms decreased from 3.00 (1.00–4.00) to 2.00 (1.00–4.00) in both treatment groups (P < 0.001 for changes within group, Wilcoxon test). The median (range) conjunctivitis scores for severity of symptoms decreased from 2.00 (0.00–3.00) to 1.00 (0.00–3.00) in both treatment groups (P < 0.001 for changes within group, Wilcoxon test). The median (range) asthma scores for frequency of symptoms decreased from 3.00 (1.00–4.00) to 2.00 (1.00–4.00) in both treatment groups (P < 0.001 for changes within group, Wilcoxon test). The median (range) total rhinoconjunctivitis score decreased from 5.00 (0.00–6.00) to 2.00 (0.00–6.00) in the SLIT group and from 5.00 (2.00–6.00) to 2.00 (0.00–6.00) in the SCIT group (P < 0.001 for changes within group, Wilcoxon test). Mean values are shown in Figure 1. No significant differences were detected between SLIT and SCIT for changes in any of the symptom scores (Mann–Whitney U test).

Figure 1. Mean rhinoconjunctivitis score in patients with SCIT or SLIT.

Differences between groups were compared by analysis of variance (ANOVA) with significance assumed for P < 0.05. We found no significant effect of age, gender, duration of rhinitis, sensitization (mono or poly), base-line scores, pollen extract, or asthmatic status between the two treatment groups on the dependent variables rhinoconjunctivitis score, rhinitis or conjunctivitis score in the different ANOVA models.

Similar improvements in the symptom scores were determined in patients with asthma versus patients without asthma (Figure 2), treatment with tree pollen extract versus treatment with grass pollen extract (Figure 3), SLIT versus SCIT in the two subgroups of patients without and with asthma, the various age groups, patients with various durations of disease, and SLIT versus SCIT in the two subgroups of patients treated with tree pollen extract versus patients treated with grass pollen extract.

Figure 2. Mean rhinoconjunctivitis score in patients with and without asthma (pooled data from all studies).

Figure 3. Mean rhinoconjunctivitis score in patients with grass or tree pollen extract (pooled data from all studies).

The general tolerability of both treatment forms was good, with mostly mild to moderate local symptoms in the oral cavity (SLIT) or at the injection site (SCIT). However, two anaphylactic reactions and serious adverse events were reported with SCIT, in contrast to none with SLIT.

Discussion

This individual patient data (IPD) meta-analysis of three open, prospective observational studies on high-dose SLIT and one open, prospective observational study on SCIT using standardized grass and tree pollen allergen extracts (Staloral 300^® and ALUSTAL^®) demonstrated that the effectiveness of SLIT was the same as that of SCIT in patients with allergic rhinitis due to grass and/or tree pollens in the daily medical practice setting.

The recommended treatment period for SIT is 3 years (1). It was possible to follow the patients for this time period only in our study with SCIT. The symptom score after 3 years was further reduced to 50% (data not shown) which was in line with improvements in a randomized controlled trial (RCT) with grass pollen SCIT (15). Three years’ data from RCT with grass pollen SLIT drops (16) and tablets (17) also demonstrate an increasing efficacy year by year. Thus, we are confident that the same effectiveness can be expected for the patients with SLIT who continued the treatment beyond the observation period.

Due to the non-interventional character of the trials, a placebo treatment was not foreseen in the studies. However, the ‘placebo effect’ is usually a rescue medication effect or an effect of fewer symptoms during lack of exposure (18). The results of our analyses are in line with a randomized, placebo-controlled, double-blind, double-dummy study including 71 adult patients with allergic rhinitis to birch pollens treated for two consecutive years after a base-line year (10), a further even smaller randomized study in 47 adult patients with allergic rhinitis to Betulaceae (alder, birch, and hazel) pollens (11), a randomized, placebo-controlled, double-blind study in 213 patients with allergic rhinitis to grass pollens (16), and also with other observational studies on SLIT in children and adults (19–22). The effectiveness of SLIT in daily medical practice compares to the efficacy seen in controlled clinical trials (4,23). Efficacy of sublingual as well as subcutaneous immunotherapy was confirmed in two recent Cochrane analyses (7,24). The standardized mean differences (SMD) for the symptom and the medication score in these publications differ numerically in favor of SCIT, but the overlapping confidence intervals do not indicate a difference between treatment modalities.

The trials in our IPD meta-analysis were prospective observational studies with a retrospective assessment of the symptoms in the pollen season before the treatment and observation period. This approach was chosen for practical reasons and was considered as robust, because the retrospective assessment of seasonal symptoms has been evaluated in a study with birch-pollen-allergic patients (25). Symptoms have been assessed by the same four-point verbal descriptor scale as was used in our studies. The retrospective assessment could detect the impact of SIT on nasal and eye symptoms. The authors concluded that retrospective assessment of symptoms tends to over-rate average symptom severity but appears sufficiently sensitive to detect treatment efficacy.

We performed a two-step IPD meta-analysis of four large observational studies because this method is the preferable approach to account for the correlations between repeated observations in the analysis of several studies (26). This approach allows for the possibility of study by treatment and study by treatment by time interactions. Missing data are assumed to be missing by random, both at the patient and at the study level. It has been demonstrated that a two-step IPD meta-analysis of several studies reveals more reliable results based on larger patient numbers than those usually achieved in double-blind randomized studies (26).

Randomized controlled trials (RCTs) with placebo and an active comparator are designed to assess the safety and efficacy of drugs, and the design of RCTs emphasizes internal validity over generalizability (27). The active comparison with a control group was not possible in our study because of the non-interventional character of observational studies, but comparisons with placebo are not necessary in meta-analyses, and head-to-head evaluations are getting more common to compare the effectiveness of different treatments.

Effectiveness studies, in which treatments are studied under real-world conditions, usually account more for generalizability to actual users, but data may be biased in other ways (27). As there were no confounding factors identified by the ANOVAs performed in our IPD meta-analysis, we trust in the validity of our effectiveness assessments.

In conclusion, following a robust meta-analysis of the data of 847 patients with allergic rhinitis due to grass and/or tree pollens, the effectiveness of SLIT and SCIT with standardized grass and tree pollen extracts appears virtually equal in daily medical routine. However, due to the advantageous safety profile and the lower level of burden within the patients’ daily life, SLIT may be considered as superior to SCIT.

Acknowledgements

Jochen Sieber and Kija Shah-Hosseini contributed equally to this manuscript.

Declaration of interest: Jochen Sieber is an employee of Stallergenes GmbH. Kija Shah-Hosseini declares no conflict of interest. Ralph Mösges has been a speaker or consultant or has received research or other funding from several pharmaceutical companies in addition to Stallergenes.