Review question(s)

What is the efficacy of antigen-specific therapeutic approaches in adult patients with MS?

Are the antigen-specific tolerance-inducing therapeutic approaches safe in adult patients with MS?

Inclusion criteria

Search strategy

The search strategy aimed to locate both published and unpublished studies. An initial limited search of PubMed, Web of Science and Cochrane was undertaken to identify articles on the topic. The text words contained in the titles and abstracts of relevant articles, and the index terms used to describe the articles were used to develop a full search strategy for PubMed (see S1 Appendix in the Supporting Information). The search strategy, including all identified keywords and index terms, was adapted for each included database and/or information source. The reference list of all included sources of evidence was screened for additional studies. Only studies published in English language were included without any date limitation.

The exact dates of the initial searches are given in S1 Appendix in Supporting Information, but as we did not finish our review in a reasonable time, we performed the search again on 5th February 2024. Results for these searches are given in S2 Appendix in the Supporting information.

Information sources

The databases to be searched included medical literature: MEDLINE via PubMed, Web of Science, The Cochrane Library, CINAHL, Google Scholar, and websites of clinical trial registrations, including ClinicalTrials.gov, WHO International Clinical Trials Registry Platform. To identify additional studies, a manual search for relevant review articles and reference lists of included articles was also performed.

Study selection

Following the search, all identified citations were collated and uploaded into EndNote X9 (Clarivate Analytics, PA, USA) and duplicates were removed. Following a pilot test, titles and abstracts were screened by two independent reviewers (H.K.Ö. and D.M.) for assessment against the inclusion criteria for the review. Potentially relevant studies were retrieved in full, and their citation details imported into the JBI System for the Unified Management, Assessment and Review of Information (JBI SUMARI) (JBI, Adelaide, Australia) [31]. The full text of selected citations was assessed in detail against the inclusion criteria by two independent reviewers (H.K.Ö. and D.M.). Reasons for exclusion of papers at full text that do not meet the inclusion criteria were recorded and reported in the systematic review. Any disagreements that arose between the reviewers at each stage of the selection process were resolved through discussion, or with a third reviewer (O.S.). The results of the search and the study inclusion process was reported in full in the final systematic review and presented in a Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram [32] (Fig 1).

Assessment of methodological quality

Eligible studies were critically appraised by two independent reviewers (H.K.Ö. and D.M.) at the study level for methodological quality in the review using standardized critical appraisal instruments from the Joanna Briggs Institute for experimental and quasi-experimental studies. Authors of the papers were contacted to request missing or additional data for clarification, where required. Any disagreements that arose were resolved through discussion, or with a third reviewer (O.S.). Given the limited quantity of expected literature in this field, studies were not excluded based on low methodological quality and high risk of bias; rather, study quality was considered when analyzing and interpreting results.

Data extraction and synthesis

Data were extracted from studies included in the review by two independent reviewers (D.M. and H.K.Ö.) using the standardized data extraction tool. The data extracted included specific details about the populations (e.g., number, age, group and type of MS) study methods (e.g., RCT, non-RCT), interventions (e.g., DNA-based, cell-based, protein/peptide- based), how the study was sponsored, and outcomes (reported clinical outcome measures of efficacy and safety such as AEs, SAEs, MRI results) of significance to the review objective. Any disagreements that arose between the reviewers were resolved through discussion, or with a third reviewer (O.S.). Authors of papers were contacted to request missing or additional data, where required.

The efficacy outcomes (ARR, MRI data and EDSS scores) were presented in a highly inconsistent manner and were often missing. Namely, some studies reported mean values with or without standard deviation, others reported medians; certain studies reported only number of patients showing clinical improvement, etc. Therefore, it was not possible to perform meta-analysis for the efficacy outcomes using either categorical or continuous data.

Data for safety were categorical data, which were described as numbers or percentages. For AEs in all studies, including both RCTs and non-RCTs, overall proportions and 95% confidence intervals were calculated by meta-analysis for proportions. This was not possible for SAEs, because most of the studies did not report any SAEs. RCT were pooled in comparative meta-analysis using the Mantel-Haenszel method. In cases of 0 events in one arm of trial, a fixed value of 0.5 was added to all cells of study results. Effect sizes were expressed as risks ratios (RR) and their 95% confidence intervals. Heterogeneity was assessed statistically using the standard chi squared and I squared tests. Statistical analyses were performed using fixed- and random-effects models for smaller (less than 5) and higher number of studies (more than 5), respectively [33]. Where statistical pooling was not possible, the findings were presented in narrative form including tables and figures to aid in data presentation where appropriate. For visualization of the results, forest plots were used.

Meta-analyses were performed separately for therapies based on DNA, protein/peptide and cells. In certain cases, there were large differences in study sample sizes within these groups, so additional subgroup analyses were performed based on sample size and quality of studies (critical appraisal score>50%). Consequently, sensitivity analyses were conducted to assess whether sample size or quality of studies affected the results of the meta-analyses.

To minimize the risk of bias due to missing results in a synthesis, particularly from reporting biases in meta-analysis, we have looked for selective reporting: we checked the primary studies for discrepancies between protocols (where possible, from ClinicalTrials.gov) and reported outcomes in the publications. Missing outcomes that were pre-specified or inconsistently reported outcomes across studies may signal selective reporting.

Assessing certainty in the findings

The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach for grading the certainty of evidence was followed and a Summary of Findings (SoF) was created using GRADEPro-GDT [34]. The SoF presented the following information where appropriate: absolute risks for the treatment and control, estimates of relative risk, and a ranking of the quality of the evidence based on the risk of bias, directness, heterogeneity, precision and risk of publication bias of the review results. The safety outcomes reported in the SoF were frequencies of the AEs and SAEs in DNA-, protein- and T cells-based therapies.

Studies inclusion

The results of the search are shown in the Prism diagram (Fig 1) and in S1 Table in the Supporting Information. The search yielded 2644 results and in total 26 studies were included in the final analysis. Twelve studies were RCTs and 14 were quasi-experimental. For studies which were still ongoing or were finished, but the results were still not published, we have contacted the researchers and asked for an update on the studies. For a study by Zubizarreta, et al. 2019. [21] ewe have received additional data for 5 patients (in total 13 patients). Moreover, for a still ongoing phase 1/2 study to evaluate the safety and efficacy of ATA188 in subjects with progressive multiple sclerosis (ClinicalTrials.gov Identifier: NCT03283826), we have received preliminary data from the authors and these data were also included in the analysis. Finally, publication by Chataway et al., 2018 [29]. Included 2 quasi-randomized studies, so we have analyzed these independently (designated as Chataway 2018a and Chataway 2018b).

The updated search did not find any relevant result, with the exception of a conference paper which presented updated findings from the study mentioned above (Phase 1/2 study to evaluate the safety and efficacy of ATA188 in subjects with progressive multiple sclerosis (ClinicalTrials.gov Identifier: NCT03283826).

Discontinued studies

There was only one study that was discontinued early, due to high frequency of allergic reactions [35]. However, a considerable number of studies were still officially ongoing, despite the estimated completion date having passed, or had been completed, yet their results had not been posted (S1 Table in the Supporting Information). We reached out to all the responsible researchers regarding the studies’ outcome, but unfortunately, we did not receive any response.

Studies and subjects’ characteristics

All reports included data mainly from phase I or phase II clinical trials, including 6 dose-escalating studies. There was only one phase III clinical trial [36]. In total, there were 1427 subjects from the RCTs, and 314 from non-RCTs. Sample size of studies ranged from 10 to 612 patients. From RCTs, most of the studies involved subjects suffering from RRMS and SPMS [37–39], RRMS or clinically isolated syndrome (CIS) [40], RRMS only [35,41–43], progressive MS [44], RPMS only [45], SPMS only [36,46]; from non-RCTs, studies included subjects with RRMS [29,47,48], RRMS or SPMS [49–51], progressive MS [52–54] or any type of MS [21,55,56].

Detailed characteristics of the subjects included in the final analysis are presented in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of included studies - randomized controlled trial form.

https://doi.org/10.1371/journal.pone.0320814.t001

The majority of the studies were funded by a profit manufacturer (N = 15), some reported an academic sponsor/research center (N = 7), and others did not report the sponsor explicitly (N = 6). Studies were conducted in Europe, United States of America, Canada, Israel, and Australia.

Therapies

In 12 RCTs, therapy consisted of products based on T cells (2 trials), DNA-based technology (2 trials), and peptide-based products (8 trials). Both trials regarding DNA – based technology studied BHT 3009, which is a MBP-coding gene [37,41]; both T cell-based vaccines evaluated attenuated T cells specific for myelin peptides, such as MBP, MOG or PLP [40,45]. Targeted peptides in the trials were TCR peptides [38], TCR ligand [39], MBP [35,36,44], complex of DR2 and MBP [46], mixture of MBP, PLP and MOG [43], and alpha B-crystallin [42]. In all trials, placebo was the control (in the study by Bourdette et al. 2005 [38], there were 3 groups: placebo, peptides dissolved in incomplete Freund’s adjuvant (IFA) and the same peptides administered without IFA). Administration route included intramuscular (i.m.) [37,38,41], intravenous (i.v.) [36,39,42,44,46], subcutaneous (s.c.) [35,40,45], and transdermal (t.d.) application [43]. The length of treatment varied highly from one day (single i.v. infusion) to 84 months.

Regarding non-RCTs, there were 8 T cell-based studies and 6 studies evaluating peptide-based therapies. Treatments were administered s.c. [29,47,49–51,57], i.m. [55,56], p.o. [58], intradermally (i.d.) [29,52], and i.v [21,48,53,54]. Treatments including T cells targeted Epstein-Barr virus (EBV) [53,54], whole myelin [57], myelin peptides MOG, PLP, MBP [21,50], or MBP and MOG [47], MBP only [51] or included T regulatory (Treg) cells [48]. From peptides, regions of MBP [29,59], peptides from V region of TCR [52,55,56], and bovine myelin [58] were assessed. Length of treatment varied from 4 weeks up to 5 years.

Critical appraisal results – quality assessment

In total, 26 studies were critically appraised for methodology quality using either the quasi-experimental studies or the RCT critical appraisal checklist, appropriately to the study design (please see Figs 2 and 3 for full methodological quality assessment). Critical appraisal scores for the RCTs were in the range 31% to 92% (Fig 2). Most important methodological concerns were associated with true randomization and treatment-blinding for participants, as only 25% of the critically appraised studies had satisfactory appraisal for these requirements. Therefore, most of the studies were at a moderate/severe risk of bias.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flowchart diagram presenting the selection of eligible studies.

https://doi.org/10.1371/journal.pone.0320814.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Critical appraisal of randomized controlled trials.

Y = yes; N = no; U = uncertain. JBI critical appraisal checklist for randomized controlled trials: Q1: Was true randomization used for assignment of participants to treatment groups? Q2: Was allocation to treatment groups concealed? Q3: Were treatment groups similar at baseline? Q4: Were participants blind to treatment assignment? Q5: Were those delivering treatment blind to treatment assignment? Q6: Were outcome assessors blind to treatment assignment? Q7: Were treatment groups treated identically other than the intervention of interest? Q8: Was follow-up complete, and if not, were strategies to address incomplete follow-up utilized? Q9: Were participants analyzed in the groups to which they were randomized? Q10: Were outcomes measured in the same way for treatment groups? Q11: Were outcomes measured in a reliable way? Q12: Was appropriate statistical analysis used? Q13: Was the trial design appropriate, and any deviations from the standard RCT design (individual randomization, parallel groups) accounted for in the conduct and analysis of the trial.

https://doi.org/10.1371/journal.pone.0320814.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Critical appraisal of quasi-experimental studies.

Y = yes; N = no; U = uncertain; Y = yes; N = no; U = unclear. JBI critical appraisal checklist for quasi-experimental studies (non-randomized experimental studies): Q1: Is it clear in the study what is the “cause” and what is the “effect” (i.e., there is no confusion about which variable comes first)? Q2: Were the participants included in any comparisons similar? Q3: Were the participants included in any comparisons receiving similar treatment/care, other than the exposure or intervention of interest? Q4: Was there a control group? Q5: Were there multiple measurements of the outcome both pre and post the intervention/exposure? Q6: Was follow up complete and if not, were differences between groups in terms of their follow up adequately described and analyzed? Q7: Were the outcomes of participants included in any comparisons measured in the same way? Q8: Were outcomes measured in a reliable way? Q9: Was appropriate statistical analysis used? *Ongoing study.

https://doi.org/10.1371/journal.pone.0320814.g003

For quasi-experimental studies, critical appraisal scores were in the range 45% to 78% (Fig 3). All studies included pre-post design. The main source of bias were the lack of control group and multiple measurements of the outcome both pre and post the intervention/exposure in all studies. None of the studies was positively appraised for these requirements. Moreover, in 73% of the studies follow up was incomplete, mostly because the authors failed to report the reasons for loss to follow up. Another confounding factor, which was present in approximately half of the studies, was the simultaneous use of immunosuppressive therapies in participants.

Efficacy outcomes

Efficacy outcomes are presented in Tables 1 and 2. Most of the studies were phase I or II, so examination of the treatment efficacy was not their primary goal. Additionally, as previously mentioned, the results of these outcomes were highly inconsistently reported or missing. In both RCTs and quasi-experimental trials, efficacy of the treatments was measured by established scales, ARR and MRI data. Most frequently used scale was EDSS, often accompanied with other scales, such as MSFC, MSIS29, 9HPT, T25-FW, and PASAT. Moreover, very different endpoints of MRI were followed and reported in these studies: some studies looked for new Gd+  T1-weighted and T2-weighted brain MRI scans, others looked for their volumes, certain authors reported their cumulative values before and after the treatment, while others looked at the fraction of patients with active scans before and after the treatment. Therefore, the efficacy of all antigen-specific treatments could not be systematically analyzed and remains ambiguous. In the following section, we narratively present the most important results from the analyzed studies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics of included quasi-experimental studies.

https://doi.org/10.1371/journal.pone.0320814.t002

Two RCTs examined the same DNA-based treatment (BHT-3009, a myelin basic protein-encoding plasmid) [37,41]. Studies reported similar results; they observed some positive trends in the several MRI endpoints, including a reduction of occurrence rate of new Gd+ enhancing lesions. However, there was no improvement in ARR, as well as in EDSS scores.

Regarding peptides, none of the RCTs showed any significant clinical benefit quantifying the disability in MS by scales. However, there was one study which examined the treatment with MBP8298 and found it caused a significant delay of EDSS progression in a major subgroup of patients with HLA haplotypes DR2 and/or DR4 [44]. This finding highlighted the importance of genetic background for the potential treatment success. However, this observation was not confirmed in a larger phase III RCT [36]. Regarding MRI results, some studies reported certain favorable findings only in the treatment group, whereas some studies did not report differences between the placebo and the treatment group [36,46]. In the trial examining TCR peptides (BV5S2, BV6S5 and BV13S1), the immunological responders tended to exert less disease activity (scan was considered to be active if there were one or more enhancing lesions) by week 24, although the differences did not achieve statistical significance. Moreover, the treatment with a small heat shock protein alpha B-crystallin (HspB5) resulted in a progressive decline in MS lesion activity, i.e., in a 76% reduction in the number and total volumes of active MRI lesions at 9 months [42]. Finally, in a small study, a one-year t.d. application of the myelin peptide skin patch showed a 66.5% reduction in the cumulative number of active lesions [43]. This study also reported a decline in the ARR in the treatment group only. Regarding quasi-experimental studies examining peptides-based therapies, it is difficult to draw a conclusion in terms of efficacy, as all studies included one-group pretest-posttest design with a small number of patients and mostly intended to assess safety of the therapies (for more details, see Table 2). From these studies, maybe the most distinguished one was examining ATX-MS-1467 (cocktail of epitopes from 4 regions of MBP) [29]. An i.d. route of administration was more favorable than the s.c. route in this study. It resulted in a significant decrease in the number of T1 Gd+ lesions at the end of the study (20 weeks). There was no change in EDSS and MSFC scores, but PASAT scores were improved.

There were 2 RCTs investigating therapies including specific T cells: one examining vaccine with T cells (TCV) against 9 antigens which were sequences of MBP, MOG or PLP [45], while the other looked into the effects of Tovaxin®, the vaccine including in vitro expanded myelin-reactive T-cells against MBP, MOG or PLP [40]. In both studies therapies did not affect MRI endpoints, but some beneficial clinical outcomes were reported. There was a significant decrease in EDSS scores and in 10-meter walking time in the TCV group [45]. Significantly more patients in the TCV group remained relapse-free compared to the placebo group. Treatment with Tovaxin® did not display any clinical benefit compared to placebo in terms of ARR and EDSS. However, a post-hoc analysis was performed using data from patients with more active baseline disease (ARR ≥ 1, ARR > 1), all of whom were naïve to DMT before joining the study. This analysis revealed that these patients experienced greater improvement in disability measured by EDSS. A previous smaller quasi-experimental study with Tovaxin®, where patients with ARR > 1 were included, showed similar results [50]. The quality of quasi-experimental studies examining T-cells-based therapies, was similar to the quality of those investigating peptide-base therapies, or slightly better (except for [48]). It is comparably difficult to draw a conclusion with respect to the efficacy of T-cells-based therapies, as all studies included one-group pretest-posttest design with a small number of patients (for more details, see Table 2). It is, however, worth mentioning the novel approach of adoptive T cell therapy against EBV, which has come up as a potential therapeutic strategy for the treatment of MS. Two studies (one is still ongoing) investigating this kind of therapy were included in this analysis [53,54]. In the first study [53], there was a symptomatic improvement (including reducing fatigue and in some cases decline in EDSS score) in 7 of 10 patients. The number of active lesions on MRI increased in 3 patients, but interestingly, all these patients showed symptomatic improvement. In the ongoing study [54], 9 from 24 subjects have reached sustained disability improvement (SDI) which was based on the improvement in EDSS scores or decrease in the time needed to walk 25 feet. Moreover, 13 participants displayed stable EDSS scores and 4 subjects perceived disability progression. After 4 years of follow-up, 5 patients preserved clinical improvements with a median duration of improvement of 27.5 months. Additionally, 8 subjects with stable EDSS scores also remained stable for a median of 41.2 months. Importantly, these patients also showed evidence of increased levels of a biomarker for myelin density, as indicated by magnetization transfer ratios (MTR) on MRI scans. This was a phase I of the clinical trial NCT03283826 and the second part is still ongoing (II phase RCT).

Safety outcomes

The monitored safety outcomes and the observed main results of the studies are presented in Tables 1 and 2. Specific AEs and SAEs reported in the studies are presented in detail in Table 3. Description of AEs and SAEs in both RCTs and quasi-experimental studies was highly heterogeneous; some studies reported both related and unrelated AEs and SAEs, whereas others stated only that there were no AEs/SAEs related to the treatment; some studies reported specific AEs and SAEs, while the others reported the organ system related to AE/SAE). Unfortunately, only several studies reported the grade of severity of AEs symptoms (mild, moderate, severe). Therefore, we were unable to group AEs into different grades (mild, moderate, severe).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Safety data reported in studies.

https://doi.org/10.1371/journal.pone.0320814.t003

For RCTs, comparative meta-analyses were performed for both safety outcomes (AEs and SAEs) and included 1418 patients (Figs 4 and 5). For all three types of treatment, there was no statistical difference in occurrence of AEs between the treatment- and placebo-related AEs (for DNA-based treatment RR was 1.06 (0.94–1.10), p = 0.334; for peptides-base treatments RR was 1.04 (0.90–1.08), p = 0.115; for T-cells-based treatments RR was 1.31 (0.97–1.76), p = 0.08), (Fig 4). Heterogeneity for these analyses was moderate (I² was 40–50%), except for T cells-base therapies, where it was low (I² was 23%). The heterogeneity for DNA-based studies was probably mainly due to differences in samples size of the only two studies (14 subjects in Barr-Or et al. [37] and 165 subjects in Garren et al. [41]). Sensitivity analysis was carried out for peptide-based studies based on sample size and critical appraisal score (studies having<50% were excluded). Analysis which included only the small sample or bigger sample size studies exerted higher heterogeneity (results not presented). When a study with low critical appraisal scores was excluded, heterogeneity was slightly decreased (I² was 50%), whereas the results remained similar to the original analysis: RR was 1.03 (0.99–1.08), p = 0.134 (Fig 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Meta-analysis of the frequency of adverse effects reported in randomized clinical trials.

https://doi.org/10.1371/journal.pone.0320814.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Sensitivity analysis in meta-analysis of the frequency of adverse effects in the peptide-based treatments when a study with low critical appraisal score was omitted.

https://doi.org/10.1371/journal.pone.0320814.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Proportional meta-analysis of the frequency of adverse effects reported in all studies (both randomized clinical trials and quasi-experimental studies).

https://doi.org/10.1371/journal.pone.0320814.g006

Proportional meta-analysis was executed only for AEs, as there were no reported SAEs in quasi-experimental trials. In total it included 1418 patients from RCTs and 203 from non-RCTs (Fig 7). Tolerogenic dendritic cell-based study (Zubizaretta et al. 2019) [21] was not included in the analysis of AEs in T cell therapies, as the mechanism of this approach is quite different from T cell-based therapies. The occurrence of at least one AEs in DNA-based therapies was 0.83 (0.776–0.879) and was lower for peptide- and T-cell-based treatments: 0.579 (0.346–0.797) and 0.535 (0.236, 0.822), respectively. Almost all studies reported injection site reactions, including redness, pain, and irritation (for more details, see Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Meta-analysis of the frequency of serious adverse effects reported in randomized clinical trials. Note: study by Freedman et al, 2011. has reported all SAEs, but did not distinguish those which were considered to be related to the treatment from those who were not.

https://doi.org/10.1371/journal.pone.0320814.g007

SAEs were reported in a particularly vague manner. Four RCTs reported SAEs. Kappos et al. reported a case of clinical exacerbation [35]; Freedman et al. [36] reported the total number of SAEs (41 in the treatment and 51 in the placebo group), and then stated that the majority of SAEs were not related to the treatment, these SAEs were also not specified; Fox et al. reported 5 SAEs from 5 patients in the treatment groups and 14 SAEs from 9 patients from the placebo group, however, as stated by the authors: “these events were listed as SAEs due to the need of inpatient treatment, but they were not severe in intensity or considered by principal investigators or study physicians to be related to study treatment” [40].

There were no differences in RR for SAEs between the treatments either for DNA-based treatment (RR was 0.63 (0.25–1.58), p = 0.322) or peptide-based treatment (RR was 0.86 (0.62–1.19), p = 0.361) (Fig 5). I² for these meta-analyses was 0. There were no reported SAEs for T cell-based treatments, so meta-analysis for these therapies was not performed (exception is Fox et al. [40], see the explanation above).

It is important to emphasize that the analysis of SAEs for peptides may be influenced by the large number of SAEs reported in the study by Freedman et al [36], where it was not clearly specified how many SAEs were related to the treatment. The authors have reported the total number of SAEs, and then stated that the majority of SAEs were not related to the treatment. As this was a large trial and the only III phase RCT in this review, we have decided to include it in the analysis.

There was one study which reported high occurrence of allergic reactions (9%) and was prematurely suspended [35].

The secondary outcome of this review was to determine the number/proportion of withdrawals from the study due to AEs/SAEs. This is presented in detail in Table 3 for every study. It seems there was a small number of withdrawals from studies, with the exception of the study which was discontinued due to hypersensitivity reactions [35].

We did not find any discrepancies between the protocols (from ClinicalTrials.gov) and reported outcomes in the publications.

SoF according to GRADE approach are presented in Table 4.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Adverse effects and serious adverse effects in antigen-specific therapies vs. placebo.

https://doi.org/10.1371/journal.pone.0320814.t004