Immunogenicity and safety of routine vaccines in children and adolescents with rheumatic diseases on immunosuppressive treatment — a systematic review

The immunogenicity of vaccines in children with juvenile autoimmune rheumatic diseases (JARDs) can be reduced, there are additional safety concerns around vaccination, and there is a potential for worsening in disease activity. In this systematic review, we summarise studies that investigated the immunogenicity and safety of routine vaccines in children and adolescents with JARD on immunosuppressive treatment. We identified 37 studies investigating 2571 children and adolescents with JARD on immunosuppressive treatment and 4895 control children. Of the 56 geometric mean antibody titres measured, 19 (34%) were lower, six (11%) higher, and 31 (55%) similar; of the 39 seroprotection rates measured, 10 (26%) were lower, two (5%) higher, and 27 (69%) similar; and of the 27 seroconversion rates measured, nine (33%) were lower, two (8%) higher, and 16 (59%) similar in children with JARD on immunosuppressive treatment compared with control children. However, many of the studies were underpowered, and not designed to show non-inferiority between children with JARD and controls. Subgroup analysis for different types of immunosuppressive treatments was not feasible, as most studies did not report results by treatment. Severe adverse events were reported in 38 children (33 with juvenile idiopathic arthritis, four with systemic lupus erythematosus, and one in a healthy child); most of them were likely not related to the vaccination (e.g. elective hospitalisation or surgery). A worsening in disease activity was reported in 44 (2%) children with JARD; again, many of them were likely not related to the vaccination. There were no safety concerns with live attenuated vaccines; however, only few studies reported results for this. Conclusion: Vaccination in children with JARD on immunosuppressive treatment is safe and should be promoted, especially since these children are at increased risk for infection. The importance for the completion of vaccination schedules should be stressed. Strategies to compensate for the lower vaccine responses, which are found in approximately one-third of these children, include measuring antibody levels to determine the optimal timing for the administration of additional booster doses. What is Known: • Children with juvenile autoimmune rheumatic diseases (JARDs) are at higher risk for infections, due to their underlying disease and their immunosuppressive treatment. • In children with JARD, the immunogenicity of vaccines might be reduced, and concerns about safety or the potential for worsening in disease activity after vaccination exist. What is New: • Our systematic review shows that vaccines in children with JARDs on immunosuppressive treatment are safe and immunogenic. • There are several limitations of the currently published studies, including random timing of measuring vaccine responses and age differences between children with JARD and control groups. Many of the studies were underpowered, and not designed to show non-inferiority between children with JARD and controls. Supplementary Information The online version contains supplementary material available at 10.1007/s00431-021-04283-w.


Introduction
Juvenile autoimmune rheumatic diseases (JARDs) are frequent in children and adolescents. The global incidence of juvenile idiopathic arthritis (JIA), for example, is estimated at 1 per 1000 children and has been reported to be increasing over the past decades [1,2]. Children with JARD are at higher risk for infection due to their underlying disease but also because they are often on immunosuppressive treatment. Vaccination is the most effective and economic method of preventing infectious diseases [3]. Children with JARD are often under-vaccinated and therefore at higher risk for vaccine-preventable diseases, as parents and paediatricians may refuse or delay vaccinations due to safety concerns [4,5]. Furthermore, compared to healthy children, the immunogenicity of vaccines in children and adolescents with JARD can be reduced and there might be concerns for a potential in worsening in disease activity [6,7].
The immunogenicity and safety of vaccines in children with JARD have previously been reviewed [8,9]. However, in the past decade, new immunosuppressive agents have become available for the treatment of JARD in children and additional studies have been published on the immunogenicity of vaccines in this group. Therefore, there is a need of an updated overview on this topic to assure the safe and most beneficial use of vaccines in these children.
In this systematic review, we summarise studies that have investigated the immunogenicity (humoral responses) and safety of vaccines in children and adolescents with JARD on immunosuppressive treatment.

Systematic review methods
A systematic search was done according to the preferred reporting items for systematic reviews and meta-analyses, the PRISMA guidelines [10]. In March 2021, MED-LINE (1946 to present) and Embase (1947 to present) were searched using the Ovid interface with the following search term combination: "child" AND "vaccination" AND "immunosuppressive treatment" (see supplementary data for detailed search terms). No language limitations were used. We included original studies which investigated the immunogenicity (specific immunoglobulin G responses) and safety of routine vaccines in children and adolescents up to the age of 21 years with JARD on immunosuppressive treatment. Exclusion criteria were studies which (i) did not specify the immunosuppressive treatment, (ii) included children with renal insufficiency or on dialysis, (iii) had less than 10 participants, and (iv) did not report results for children separately to those from adults. References of retrieved articles were hand-searched for additional publications.
The following variables were extracted from the included studies: author, publication year, country, study type, level of evidence, number of participants, age and gender of participants, immunosuppressive treatment, vaccine type, vaccine brand, vaccine producer, vaccine dose, number of vaccine doses, interval between doses, timing of blood sampling after last vaccination, antibody responses, safety (including local and systemic reactions, serious adverse events and worsening in disease activity), and additional important findings. The ROBINS-1 tool was used to assess risk of bias [11].

Systematic review results
Our search identified 3488 studies. Of these, 30 fulfilled the inclusion criteria . Seven additional studies were found by hand-searching of references [42][43][44][45][46][47][48]. The selection of included studies is summarised in Fig. 1. The 37 studies (28 cohort studies, three case-control studies, three cross-sectional studies, two randomised controlled trials (RCTs) and one case series) included in this review investigated 2571 children and adolescents with JARD on immunosuppressive treatment and 4895 control children (4865 healthy children and 30 children with non-rheumatic diseases). The number of participants in each study ranged from 23 to 2576 (median 77, mean 202). Antibodies against 22 different antigens were measured. Of the studies, 30 evaluated the safety of vaccines, 25 local reactions, 26 systemic 1 3 reactions, 27 severe adverse events (SAEs), and 26 worsening in disease activity. All of the studies were done in industrialised countries: Brazil 11, Netherlands 8, Greece 4, Turkey 3, Germany 2, Japan 2, Iran 1, Italy 1, Slovenia 1, Spain 1, Sweden 1, Switzerland 1, and USA 1. The results of these studies are summarised in Tables 1 and 2  . The risk of bias summary of studies included in the review can be found in Table 3.

Measles, rubella
One study investigated the persistence of specific antibodies after two doses of measles and rubella vaccination in 41 children with enthesitis-related arthritis (ERA) 1 and 3 years after the initiation of adalimumab and compared it with 149 healthy children [43]. At both timepoints, the GMTs were lower in children with ERA for measles and rubella compared with healthy children of similar age, while there was no difference in SPRs. No difference in GMTs or SPRs was found between children on adalimumab only and children who were on additional methotrexate (MTX) or sulfasalazine. The study provided no information about the safety of the vaccines.                       Some children were receiving more than one drug 2 Time after initiation of adalimumab (mean duration between last dose and sampling in children with ERA 7.2 (SD 3.2) vs healthy children 8.4 years (SD 2.0)) 3 Summary of all reactions occurring after 1st, 2nd, and 3rd vaccination in 128 JDM and 112 controls 4 Numbers for children and controls not specified separately 5 Higher use of immunosuppressive drugs in polyarticular onset JIA compared with oligoarticular JIA (80 vs 42% (p < 0.01)) 6 Children divided into a vaccinated group without booster (3 DTP, 1 MMR dose) and a group with booster vaccination (> 3 DTP, > 1 MMR) 7 Elevated GMT before vaccination in JARD compared to controls (19.7 vs 10.8 (p < 0.01)) 8 Child with epilepsy 9 'High avidity' defined as > 60%, 'borderline avidity' as 40-59%, 'low avidity' as < 40% 1 3

Juvenile dermatomyositis
Two studies investigated the immunogenicity and safety of vaccines in children with juvenile dermatomyositis (JDM) [23,42]. The studies investigated human papilloma virus (HPV) and influenza vaccines in 72 children with JDM and 116 controls.

HPV
One study investigating the immunogenicity and safety of three doses of a HPV16/HPV18 vaccine in 42 children with JDM and 35 healthy children, 1 and 6 months after vaccination [42]. No difference in SPRs for both serotypes was detected 1 month after the last vaccination. Six months after the last vaccination, SPR for both serotypes was 94% in children with JDM. SPRs were not specified for healthy children. No difference in SPRs was found between children on different treatment regimens (steroids, hydroxychloroquine, MTX, azathioprine, cyclosporine, or mycophenolate mofetil) at either time point. No SAEs were observed. One child with JDM was reported to have a worsening in disease activity 6 months after vaccination.

Influenza
One study investigated the immunogenicity and safety of one dose of influenza vaccination (A/H1N1 strain) in 30 children with JDM and 81 healthy children, 21 days after vaccination [23]. A lower SCR was found in children with JDM. In contrast, no difference in GMT or SPR was detected. Separate results for different types of immunosuppressive treatment were not specified. No severe SAEs or worsening in diseases activity were reported.

HAV, HBV
Two studies investigated the immunogenicity and safety of two doses of HAV vaccination in 130 children with JIA and 143 healthy children [13,35]. One study found a lower GMT against HAV in children with JIA compared to healthy children 1 and 12 months after vaccination, but no difference in SCR or SPR, while the other study found a lower SPR in children with JIA 2 months after vaccination. The first study reported a worsening in disease activity in 15 (18%) children with JIA (in two after the first dose and in 13 after the second dose after a mean of 8 months) [13]. None of them had a worsening in disease activity during the first 3 months after vaccination. No SAEs were reported in the two studies. The second study investigated the immunogenicity of three doses of HBV vaccine in 39 children with JIA and 41 healthy children and found a lower GMT and SPR in children with JIA [16]. No difference in GMTs was found between children on steroids and those on MTX. No SAEs were reported.

HPV
One study investigated the immunogenicity and safety of three doses of a HPV16/18 vaccine in 68 children with JIA and 55 healthy controls [44]. No difference in GMT was found. No difference in GMTs or antibody avidity was found between children on MTX and those on anti-tumournecrosis-factor (TNF)-alpha blockers, anti-interleukin (IL)-1 blockers, leflunomide, and mycophenolate mofetil. One child on MTX did not reach seroprotection. No worsening in disease activitiy was reported. In 14 children with JIA and one healthy child SAEs were reported, many of them were elective hospitalisations or surgeries (for details see Table 2).

Diphtheria, tetanus
Two studies assessed the immunogenicity of diphtheria and tetanus vaccination in 429 children with JIA and 2176 healthy children [30,41]. One study found lower GMTs and SPRs for diphtheria and tetanus in children with JIA compared to healthy children [30]. In the other study, no control children were included, the SPR for diphtheria in children with JIA was 90%, and for tetanus 100% [41]. No difference in GMTs or SPRs was found between children on abatacept, MTX, steroids, and anti-TNF-alpha blockers [30,41]. Neither study provided information about SAEs or worsening in disease activity.

Influenza
Six studies investigated the immunogenicity and safety of one or two doses of influenza vaccines in 292 children with JIA, 154 healthy children, and 14 children with non-rheumatic diseases [12,22,28,29,31,46]. Five studies used a trivalent influenza vaccine (TIV) and one study an influenza vaccine with an A/H1N1 strain only. The GMTs were lower against at least one strain in two studies [28,31], the SCRs were lower against at least one strain in four studies [22,28,31,46], and 1 3 the SPRs were lower for at least one strain in two studies (see Tables 1 and 2) [31,46]. Most studies did not report differences in specific antibody responses between different treatment regimens [12,22,29]. However, one study reported that children on anti-TNF-alpha blockers had lower SCR and SPR against A/H1N1 compared to children on leflunomide, MTX, steroids, and cyclosporine [46]. Only one study specified SAEs; one child with JIA on etanercept needed a hospitalisation for fever and coxalgia 1 day after vaccination [31]. Eighteen children with JIA reported a worsening in disease activity 7 days to 6 months after vaccination [12,28,31].

MenC
Two studies investigated the immunogenicity of one dose of MenC vaccination in 361 children with JIA and 1527 healthy children [20,26]. One study found no difference in GMT between children with JIA and healthy ones [20]. The other study, which compared different immunosuppressive treatments in children with JIA, found a lower GMT against MenC in children treated with MTX, sulfasalazine, etanercept, infliximab, or cyclosporine compared to children on non-steroidal anti-inflammatory drugs or without treatment [26]. No information concerning SAEs was provided. No worsening in disease activity was reported.

Pneumococcus
One study each investigated the immunogenicity and safety of one dose of 23-valent pneumococcal polysaccharide vaccine (PPV23) and two doses of 7-valent pneumococcal conjugate vaccine (PCV7) in 27 and 63 children with JIA, respectively [19,38]. Both studies compared the antibody response to pneumococcus in children with JIA on MTX and cyclosporine with children with JIA on the former treatment plus additional adalimumab or etanercept. The study using the PPV23 vaccine found no difference in GMTs, SCRs, or SPRs for serotypes 4, 6B, 9 V, 14, 18C, 19F, and 23F between the two groups [19]. The study using the PCV7 vaccine found lower GMTs but not SPRs for serotypes 4, 14, and 23F in on adalimumab or etanercept compared to those on steroids or cyclosporine [38]. However, no difference in GMTs was found between children on adalimumab and those on etanercept [38]. One child on etanercept developed pneumococcal pneumonia (serotype not specified) requiring hospitalisation 5 months after vaccination [19]. In one child with JIA on additional adalimumab or etanercept, a worsening in disease activity was reported [38].

MMR
Three studies, including one RCT, investigated the immunogenicity of MMR vaccination in 546 children with JIA and 2198 healthy children [14,30,45]. In the RCT, both groups included children with JIA on immunosuppressive treatment who had been previously been vaccinated with MMR, but only one group was randomised to receive a MMR booster [45]. In the other two studies, either two doses of MMR or a monovalent measles vaccine followed by MMR were given [14,30].
One study reported a higher GMT against measles in children with JIA compared to healthy children [30]. However, the children with JIA were older than the healthy children; therefore, likely more of them had received an MMR booster, which was given at nine years of age. Another study reported a lower GMT against measles in children with JIA after receiving MTX for 6 months initiated more than 4 years after the second dose of MMR compared to healthy children [14]. But again the children were from different age groups. The children with JIA had a mean age of 16 years compared to the healthy children with a mean age of 11 years. Therefore, there was a larger time interval between vaccination and measuring vaccine antibody responses in children with JIA. Only one study measured SPRs for measles and reported a higher SPR in children with JIA compared to healthy children in the group vaccinated with two doses of MMR but not in the group vaccinated with a monovalent measles vaccine followed by a dose of MMR [30]. However, as mentioned above some of the healthy children, who were younger, likely did not yet receive two doses of the MMR vaccine. The RCT reported an increase in GMT in the children with JIA who received an MMR booster 12 months after the vaccine [45].
Two studies compared GMTs against mumps in children with JIA compared to healthy children [14,30]. One found a lower GMT against mumps in children with JIA, while the other did not find a difference in GMTs. The one study which measured SPRs for mumps reported no difference in children with JIA compared to the healthy controls when receiving two doses of MMR but a lower SPR for mumps in the children who received a monovalent measles vaccine followed by one dose of MMR [30]. The RCT reported an increase in GMT against mumps in the children with JIA who received an MMR booster 12 months after vaccination [45].
One study did not detect a difference in rubella specific GMTs between children with JIA and healthy children, while another study found a lower GMT in children with JIA [14,30]. The study which measured SPRs against rubella, reported a lower SPR in children with JIA compared to healthy children after one dose of monovalent measles vaccine followed by MMR but not in the children who received two doses of MMR [30]. The RCT reported a higher GMT in children with JIA receiving a MMR booster compared to children with JIA without the booster 12 months after vaccination [45].

3
No difference in GMTs against measles, mumps, or rubella was found between children on MTX and children on steroids, anti-TNF-alpha blockers, anti-IL-1 blockers, or leflunomide [30,45].
Only two of the three studies evaluated the safety of the MMR vaccine; no worsening in disease activity was reported in either of the studies [14,45]. Only the RCT reported SAEs; five children who received a booster and eleven children in the control group were reported to have SAEs [45]. Most of them were elective hospitalisation and surgeries, unlikely related to the vaccination. No disease due to infections with vaccine viruses was observed [45].

HBV
One study investigated the immunogenicity and safety after three doses of an HBV vaccine in 20 children with SLE and 24 healthy children [34]. A lower GMT was found in children with SLE, while the SPR was similar in both groups. Separate results for different types of immunosuppressive treatment were not reported. No SAEs were reported. Three children with SLE were reported to have a worsening in disease activity, one child 1 month after the first dose and the other two children 1 month after the second dose.

Tetanus
Two studies investigated the specific antibody response after five doses of tetanus vaccine in 70 children with SLE and 74 healthy children [15,36]. In one study, a lower GMT was found in children with inactive SLE (but not active SLE) compared to healthy children [36]. The other study did not find a difference in GMTs between children with SLE and healthy ones [15]. Separate results for different types of immunosuppressive treatment were not reported. During the study, four SAEs occurred in children with SLE; all were disseminated infections requiring hospital admission unlikely related to the vaccine (disseminated varicella infection, primary peritonitis, meningitis and pneumonia) [15]. No information concerning worsening in disease activity was provided in either study [15,36].

Influenza
One study investigates the immunogenicity of one dose of an A/H1N1 influenza vaccine in 118 children with SLE and 102 healthy children [21]. A lower GMT, SCR, and SPR were found in children with SLE. Children on a higher steroid dose did more frequently not seroconvert compared to those on a lower dose (18.0 vs 10.5 mg per day). No SAEs or worsening in disease activity were reported.

Measles
One retrospective study investigated the specific antibody response after measles vaccination in 30 children with SLE with different immunosuppressive treatments (chloroquine, steroids, azathioprine, cyclosporine, cyclophosphamide, and MTX) and 28 healthy children [36]. No difference in GMTs was found between children with active SLE, inactive SLE and healthy children. Separate results for different types of immunosuppressive treatment were not reported. No information concerning SAEs and worsening in disease activity was provided.

VZV
One RCT investigated the immunogenicity and safety of one dose of VZV vaccine in 28 children with SLE on either chloroquine, steroids, azathioprine, or MTX and 28 healthy children [47]. No difference in GMT for children with JIA compared to healthy ones was found 1, 6, and 12 months after vaccination. Separate results for different types of immunosuppressive treatment were not reported. Furthermore, no SAEs or worsening in disease activity were reported.

HAV, HBV
One study investigated the immunogenicity of two doses of a combined HAV and HBV vaccines in 78 children with JARD, including 71 (91%) children with JIA [17]. A positive SCR was reached for HAV and HBV in 100% and 93%, respectively. No difference in GMTs for different treatments was found. No information concerning SAEs and worsening in disease activity was provided.

HPV
One study investigated the immunogenicity of three doses of a HPV in 12 children with JARD (6 JDM, 6 SLE) and 49 healthy children [48]. No difference in GMTs between 1 3 JARD and controls were found 1 and 6 months after vaccination. One child with JDM without immunosuppressive treatment did not seroconvert. Separate results different types of immunosuppressive treatment were not reported. No information concerning SAEs was provided. One child with JDM was reported to have a worsening in disease activity 1 month after the second vaccination.

Tetanus
One study investigated the immunogenicity of more than three doses of a tetanus vaccine in 50 children with JARD (including 46 (92%) children with JIA) and 31 healthy children [40]. A lower GMT was detected in children with JARD on MTX with or without anti-TNF-alpha blockers compared to children with JARD without treatment or healthy children. No information concerning SAEs and worsening in disease activity was provided.

Influenza
Five studies investigated the immunogenicity of influenza vaccination in 430 children with JARD, 138 healthy children, and 16 children with non-rheumatic diseases [18,27,32,33,37]. Three studies used a TIV and two studies a single A/H1N1 strain vaccine. Two studies did not find differences in GMT, SCR, or SPR between children with JARD and controls [18,33]. One study found a lower GMT, SCR, and SPR for the A/H1N1 strain in children with JARD compared to healthy children [32]. In contrast in another study, a higher GMT for the B strain in children with JARD compared to healthy children was found, while there was no difference in SCR for all three strains [37]. One study reported a lower GMT in children on azathioprine, mycophenolate mofetil, and steroids compared to children on cyclosporine, leflunomide, or cyclophosphamide [32]. No SAEs were reported [18,27,32,33,37]. In one study a worsening in disease activity was reported in two children with JARD (one child with JIA, one with Takayasu arteritis) 2 weeks after vaccination [37].

Measles, rubella
One study investigated the specific antibody responses against measles and rubella after a minimum of one dose of MMR in 50 children with JARD (46 children with JIA) on MTX or MTX plus anti-TNF-alpha blockers and 31 healthy children [40]. No difference in GMTs for measles or rubella was found between children with JARD and healthy children. Children with JARD on anti-TNF-alpha blockers had a lower proportion of transition B cells compared with those without anti-TNF-alpha blockers and controls. No information concerning SAEs and worsening in disease activity was provided.

VZV
Three studies investigated the immunogenicity of one or two doses of VZV vaccination in 97 children with JARD on different immunosuppressive treatments and 36 healthy children [24,25,39]. No difference in GMT or SPR was found in any of the three studies. No SAEs were reported. Separate results for different types of immunosuppressive treatment were not reported. One study reported a worsening in disease activity in three children with JIA 4 to 6 weeks after vaccination [25].

Discussion
Our systematic review shows that vaccines in children and adolescents with JARDs on immunosuppressive treatment are safe and immunogenic. Overall, a decreased specific antibody response was reported in one-third (26-33%) of all measurements (GMT, SPR, or SCR). However, it is important to take into consideration the timing of the measurements. For example, a study, which measured antibodies to HAV/HBV vaccine in children with JARD, reported that the initial response after one dose was low, but after receiving a second dose, almost all children with JARD seroconverted [17]. This stresses the importance for the completion of vaccination schedules, especially in high-risk children, such as children with JARDs.
Furthermore, due to the lower vaccine response found in approximately one-third of children with JARDs on immunosuppressive treatment, additional booster doses can be offered to optimise vaccine efficacy in these children. Two studies showed an accelerated antibody loss in children with JIA [6,26]. In the included RCT, which randomised children with JIA to either receive or not receive a booster dose of MMR, a higher proportion of children who did not receive a booster were not seroprotected [45]. This demonstrates the importance of booster doses in children with JARDs. A pro-active approach to detect insufficient antibodies levels in children with JARDs might be a valuable tool to optimise the timing booster doses. Similarly, pre-travelling antibody measurements could be useful, for example, for HAV in children who are travelling to high incidence countries.
A further main concern in children with inflammatory or autoimmune diseases, including JARD, is that vaccines may trigger an onset or worsening in disease activity. For example, it has previously been reported that HBV vaccination might trigger the onset of an underlying inflammatory or autoimmune rheumatic disease [49] or that HPV vaccination could trigger the onset of SLE [50]. Many studies exclude children with active JARD to avoid potential worsening in disease activity [34]. Only one study included in the review, compared the immunogenicity of a measles and tetanus 1 3 vaccine in children with active and inactive SLE [36]. The study reported a higher immunogenicity of the vaccines in children with active compared to inactive SLE. However, the children in the latter group were older than the ones with active SLE, which likely explains some of this difference.
In our review, ten studies reported children with worsening in disease activity after vaccination [12,13,25,28,31,34,37,38,42,48]. However, often it could not be differentiated between the vaccines as a trigger for the worsening of disease or other possibilities causes, such as changes or non-compliance to the immunosuppressive treatment [31]. Furthermore, many of the JARDs are characterised by an intermittent and relapsing course even without triggers. As many care providers prefer to vaccinate these children in a stable phase and often defer vaccination until such a phase is reached, the chance of a relapse due to the normal course of the disease might be higher after vaccination, which can then be misinterpreted as a relapse triggered by the vaccine. Although, worsening in disease activity might be a more of a concern after live attenuated vaccines, only one study reported a worsening in disease activity in three of 39 children with JIA, 4 to 6 weeks after VZV vaccination [25]. In contrast, one of the studies reported a decrease in number of affected joints in children with JARD after VZV vaccination [39]. Another study included in the review, investigated the association of autoantibodies and disease activity after influenza vaccination in children with JIA, and did not find an association [28]. Importantly, it should be noted and communicated that vaccine-preventable diseases, for example mumps, measles, and rubella, can also trigger an activation or exacerbation of the underlying inflammatory or autoimmune JARD [51,52]. Overall, there is increasing evidence suggesting that vaccines do not induce significant worsening of underlying disease [8,53].
Not only for a worsening in disease activity, but also for SAEs, it is difficult to proof a correlation. Many of the SAEs reported in the studies included in this review seem unlikely to be correlated to vaccination (e.g. elective hospitalisations and surgeries). As an example, two children with SLE have been reported to have deterioration in renal function within 18 months after HPV vaccinations [54]. However, it is not clear if this is an adverse event due to the vaccination or rather illustrate the natural progression of the disease itself. As both cases were diagnosed with state four nephritis before vaccination, an evolution of the disease itself seems more likely. In another study, a febrile convulsion was reported as a severe adverse event after influenza vaccination in a child with JARD. However, this child was known for epilepsy [27]. In one study, a vesicular rash was reported in three of 49 children with JARD on immunosuppressive treatment with MTX and anti-TNF-alpha blockers [25]. Another study reported that rashes after VZV vaccination were not more common in children with SLE on immunosuppressive treatment compared to healthy children [47]. None of the studies reported an infection due to attenuated vaccine viruses, and all recovered promptly without treatment. These results are, however, limited by most studies investigating live attenuated vaccine not reporting systemic reactions, severe adverse events, or worsening in disease activity [14,30,36,40,43].
Unfortunately, there is no study which investigated the immunogenicity and safety of yellow fever vaccination in children with JARDs on immunosuppressive treatment. A recent review from the European League Against Rheumatism concluded that yellow fever vaccination in adults with autoinflammatory rheumatic diseases should be avoided due to the risk of vaccine-induced yellow fever [55]. As yellow fever is circulating in many parts of the world, the safety and immunogenicity of the vaccine in children with JARDs is an urgent future research topic.
Furthermore, the current severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2) pandemic also highlights the urgent need to assess the immunogenicity and safety of new vaccines, such as messenger ribonucleic acid-based (mRNA) vaccines, in this patient group. Three recent studies showed that mRNA and viral vector-based SARS-CoV-2 vaccines are immunogenic and safe (no increased side effects or induction of disease flares) in adults with rheumatic diseases on immunosuppressive treatment [56][57][58]. However, specific antibody responses were lower in adults on immunosuppressive treatment compared to healthy controls, especially those on steroids, rituximab, mycophenolate mofetil, and abatacept [56,58]. No data is currently available on the immunogenicity and safety of SARS-CoV-2 vaccine in children with rheumatic diseases on immunosuppressive treatment.
The strengths of our review are the systematic approach and the comprehensive literature search. The limitations are the heterogenicity of the included studies which precluded a meta-analysis, especially as vaccine responses were not always measured one month after vaccination (gold-standard) and sometimes the time interval between vaccination and measurement of responses even differed between cases and controls. Furthermore, the dose of the immunosuppressive treatment and the disease activity or severity was not specified in most of the studies. Many of the studies were underpowered, and not designed to show non-inferiority between children with JARD and controls; therefore, finding no difference between the groups does not imply equivalence. Moreover, we only evaluated antibody responses, as there are almost no studies which report on vaccine efficacy, and cellular or cytokine responses to vaccines.
In conclusion, vaccination in children with JARD on immunosuppressive treatment should be promoted and the importance for the completion of vaccination schedules 1 3 should be stressed. Strategies to compensate for the lower vaccine responses or faster decline of antibodies include measuring antibody levels to determine the optimal timing for the administration of additional booster doses. Further studies including children with active JARD are needed for evidence-based guidelines to vaccination in these children.
Authors' contributions MK and PZ designed the study. MK did the literature search and drafted the initial manuscript. LFP and PZ critically revised the manuscript, and all authors approved the final manuscript as submitted.

Conflict of interest The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.