Qualitative Analysis of Well Powered Studies
A qualitative analysis of well powered sublingual and subcutaneous immunotherapy studies was performed.
Sublingual Immunotherapy Studies
Our literature search identified six well powered trials assessing sublingual immunotherapy. These were published between 2006 and 2011. Although this review does not focus on statistical significance and efficacy of the product itself but rather on outcome measures and study design, a major search criterion was the sufficient number of participants (>100 per study arm). As all of the included studies are fairly recently published, we expected them to have been conducted in accordance with the latest guidelines and recommendations in order to present high-quality and reliable data. Whereas three studies limited the participation to children and adolescents, three studies included only adults (>17 years). Both studies published in 2011 were conducted in North America (USA and Canada), whereas the earlier studies were all performed in Europe.
As displayed in Table 1 , there is no uniform understanding of what outcome parameters should be assessed to show efficacy of SLIT. In general, three different primary outcomes were used, namely TSS, TMS and the total combined score (TCS), which is the sum of TSS and TMS.
Likewise, a great variety of secondary outcome parameters to confirm and support the study results were described: TCS, TMS, TSS for peak season, TMS for peak season, individual symptoms, overall effectiveness or global improvement, symptom–medication-free days, well days, quality of life, safety, specific IgG and IgE and pollen counts.
Despite the differences in defining the primary outcome, the assessment of symptoms is of major importance for the primary endpoint. All studies reported six identical rhinoconjunctival symptoms that need to be considered: sneezing, rhinorrhoea, nasal pruritus, nasal congestion, ocular pruritus and watery eyes. Only the wording of each symptom differed slightly with some using the medical term, whereas others a more informal expression to describe the same symptom (e.g. rhinorrhoea = runny nose). However, as guidelines suggest allergic rhinoconjunctivitis should be defined by identical symptoms among the studies.
Furthermore, a uniform understanding exists regarding the scaling system for rating each of the relevant symptoms. A standardized 4-point scale ranging from 0 for no symptoms, 1 for mild symptoms, 2 for moderate symptoms to 3 for severe symptoms was applied in all studies.
On the contrary, far less standardized was the measurement of the TMS. Two studies used apart from the TSS, also the TMS as a primary outcome parameter, whereas the North American studies took both scores and combined them in one single outcome parameter, the TCS. Of those four studies, three applied the same rescue medication with the same scoring system, whereas one study used a different scale. A third and fourth way of calculating the medication score was described when taking theTMS as a secondary endpoint. In summary, six studies used four different types of rescue medication and four different scale systems with none of these medication scores being validated yet.
Similarly disconcerting is the fact that no uniform phrasing exists with many parameters. The symptom score used as primary outcome parameter may be called 'RTSS', 'average daily symptom score' or just 'daily symptom score', all describing the same symptoms. Likewise, with 'well days' or 'symptom-free days' or 'symptom–medication-free days' used in literature, there are identical terms for different definitions or different terms for describing identical parameters.
Disappointingly, the well acknowledged and highly recognized assessment of 'quality of life' was not performed in all studies and the VAS was not applied in any of the included trials.
Although only looking at a short time span of 5 years (2006–2011) primary outcome parameters are developing towards a more complex construct. Whereas studies published in 2006/2009 only used the TSS as their primary endpoint, studies in 2009/2010 assessed both the TSS and TMS individually as primary parameters and the most recent studies in 2011 applied the recommended TCS as the primary efficacy outcome measure which shows promise for the future.
Subcutaneous Immunotherapy Studies
Despite the large number of published SCIT studies, the majority of them are carried out in small numbers of participants. Four well powered double blind, placebo-controlled, RCTs were identified from our search (Table 2). These were all carried out in adults. All except one study which was a multicentre study conducted in both North American and European centres were conducted in European countries.
Similarly to the SLIT studies, different outcome measures were used. Three of the identified studies used a combined symptom and medication score as the primary endpoint for clinical efficacy. The other study used separate symptom scores and medication use scores during the grass pollen season. Participants in all studies recorded individual symptom scores varying from 7 to 9 nasal, eye and lung symptoms. All used a 4-point scale, that is, 0 = no symptoms to 3 = severe symptoms scoring but there was variability between the number of symptoms assessed.
Furthermore, there was significant variability in the rescue medication scoring systems used. For example, a 5mg prednisolone tablet scored 2 points in one study, 4 points in another and even 12 points in another.
Two studies used a VAS as a secondary outcome; however, only one of these explained the point system used. Notably, they both included the same rhinoconjunctivitis quality of life questionnaire. However, despite using the same measure there was variability in the time-points at which this was assessed, with one study recording this before the first injection and every 2 weeks during the grass pollen season, whereas in the other study it was completed before and at the peak of the pollen season. Only one study used well and bad days measures.
Additional secondary outcomes, in the form of conjunctival provocation tests were performed in one of the studies. Specific IgE and IgG were measured in three studies.