Infections in Children and Adolescents With Juvenile Idiopathic Arthritis and Inflammatory Bowel Disease Treated With Tumor Necrosis Factor–α Inhibitors: Systematic Review of the Literature

Tumor necrosis factor alpha (TNF-α) inhibitors can increase the risk of infections. This systematic literature review describes the epidemiology, microbiology, and types of infections reported in children and adolescents with juvenile idiopathic arthritis and inflammatory bowel disease treated with TNF-α inhibitors.

Abstract

Tumor necrosis factor alpha (TNF-α) inhibitors are increasingly administered to children and adolescents with juvenile idiopathic arthritis (JIA) and pediatric inflammatory bowel disease (pIBD). Adult studies indicate that TNF-α inhibitors lead to an increased risk of serious infections compared to other disease-modifying antirheumatic drugs. We report herein a systematic literature review detailing the epidemiology and types of infections reported in children with JIA and pIBD treated with TNF-α inhibitors. The most frequently reported infections were mild and characterized as viral in etiology. Severe bacterial and fungal infections also occurred, but were less common and possibly associated with intrinsic risk factors and concurrent immunosuppressive therapy. Few pediatric patients developed Mycobacterium tuberculosis, likely due to effective screening. There were 8 infectious fatalities in children treated with TNF-α inhibitors. Overall, although rare, serious infections occur in immunocompromised children and adolescents with JIA and pIBD receiving TNF-α inhibitors.

Tumor necrosis factor alpha (TNF-α) inhibitors are increasingly being administered to children and adolescents with juvenile idiopathic arthritis (JIA) or pediatric inflammatory bowel disease (pIBD). Adult studies have, for the most part, shown that TNF-α inhibitors lead to an increased risk of opportunistic infections, serious infections, and hospitalizations, when compared to other disease-modifying antirheumatic drugs (DMARDs) [1–4]. Despite the extensive use of biologics in pediatrics, it remains unclear whether there is a similarly increased risk of infections and what types of infections occur. To date, studies in children with JIA or pIBD treated with TNF-α antagonists have consisted of small numbers of subjects, with outcomes mostly focused on drug efficacy and serious adverse events. Furthermore, strategies to prevent infections and safely monitor pediatric patients treated with TNF-α inhibitors are largely extrapolated from the adult literature. This systematic literature review describes the 3 main TNF-α inhibitors used in JIA or pIBD, summarizes the microbiology and types of infections pediatric patients treated with TNF-α inhibitors may be susceptible to, and provides an overview of strategies that may be useful for the prevention of infectious complications in this population. Characterizing the infections that occur in JIA and pIBD patients treated with TNF-α inhibitors may help prevent unnecessary treatment interruptions and unwarranted hospitalizations.

ROLE OF TNF-α IN PATHOGENESIS OF JIA AND pIBD

TNF-α is a cytokine with an important role in inflammation and immune function. It promotes an inflammatory cascade, critical in the host response to infections and local injury. However, at high concentrations, TNF-α may lead to excess inflammation and organ damage as seen in JIA and pIBD. JIA patients can have increased TNF-α levels and other proinflammatory cytokines in the peripheral blood and synovial fluid, which may lead to tissue damage, cartilage loss, and bone destruction [5, 6]. In pIBD patients, unregulated TNF-α production can exacerbate disease through its proinflammatory effects and induction of apoptosis, compromising the gastrointestinal mucosa, specifically the barrier function of the endothelial and epithelial layers [7]. TNF-α inhibitors are thought to improve the patient's clinical course by impairing the ability of TNF-α to bind to its receptors, inactivating the sepsis cascade cytokines, inhibiting production of inflammatory cytokines, and preventing further inflammation and damage to tissue or joints. However, TNF-α inhibition can potentially make the patient more susceptible to infections due to this downregulation of the immune system [8].

INFLIXIMAB, ETANERCEPT, AND ADALIMUMAB

The 3 TNF-α inhibitors most commonly evaluated in pediatrics include 2 monoclonal antibodies, infliximab (Remicade) and adalimumab (Humira) and a soluble TNF receptor, etanercept (Enbrel). All 3 can bind to soluble and transmembrane TNF, thus attenuating TNF-α's biological effects. However, they differ in structure, half-life, dosage, and indications (Table 1). Infliximab is a chimeric mouse–human monoclonal immunoglobulin G1 (IgG1) antibody. It was first approved by the Food and Drug Administration (FDA) in 1998 for adults with Crohn's disease (CD). In 2006 infliximab was approved for pediatric CD, and by 2011 it was approved for pediatric ulcerative colitis (UC). It was approved for rheumatoid arthritis (RA) in 1999, but has no FDA-approved indications in JIA at this time, although often used off-label in this population.

Table 1.

TNF-α Inhibitors Used in Juvenile Idiopathic Arthritis and Pediatric Inflammatory Bowel Disease

Biological Agent (Route Administered)	Composition	Half-life	Dosage and Timing	FDA-Approved JIA and pIBD Indications	EMEA-Approved JIA and pIBD Indications
Infliximab (Remicade) (IV)	Chimeric IgG1 monoclonal antibody	Approximately 8–9.5 d	Initial 3 mg/kg, then 3 mg/kg at 2 and 6 wk Maintenance: 3–6 mg/kg every 8 wk	Not approved in JIA	Not approved in JIA
			CD: Initial 5 mg/kg, then 5 mg/kg at 2 and 6 wk Maintenance: 5 mg/kg every 8 wk. Can increase to 10 mg/kga	Age ≥6 y with moderate to severe CD and UC	Age 6–17 y with CD and UCb
Etanercept (Enbrel) (SQ)	Human; fusion protein	Approximately 4.25 ± 1.25 d	0.8 mg/kg weekly Or 0.4 mg/kg twice weekly	Age ≥2 y with moderate to severe polyarticular JIA	Age ≥2 y with polyarticular JIAc
					Age 12–17 y with psoriatic arthritisc
					Age 12–17 y with enthesitis-related arthritisc
				No role in IBD	No role in IBD
Adalimumab (Humira) (SQ)	100% human	10–20 d	JIA:15–29 kg: 20 mg every 2 wk ≥30 kg: 40 mg every 2 wk	Age ≥4 y with moderate to severe polyarticular JIA	Age ≥4 y with polyarticular JIA
			CD: Induction: <40 kg: 80 mg day 1, 40 mg day 14 >40 kg: 160 mg day 1, 80 mg day 14 Maintenance: <40 kg: 20 mg every 2 wk >40 kg: 40 mg every 2 wk	Not approved in pIBD yet	Age 6–17 y with severe active CDb

Abbreviations: CD, Crohn's disease; EMEA, European Medicines Agency; FDA, Food and Drug Administration; IBD, inflammatory bowel disease; IgG1, immunoglobulin G1; IV, intravenous; JIA, juvenile idiopathic arthritis; pIBD, pediatric inflammatory bowel disease; SQ, subcutaneous; UC, ulcerative colitis.

^a If response is incomplete.

^b In patients who have not responded to other therapy.

^c In patients who have not responded to or cannot tolerate methotrexate.

Adalimumab, a fully humanized IgG1 monoclonal antibody, was initially approved for RA in 1998, with indications for the adult CD population by 2007. In 2008 it was approved for JIA patients age 4 and older. Recent studies have shown that adalimumab is also effective in inducing and maintaining remission in pediatric CD. In 2012 adalimumab was approved for pediatric CD in Europe, but it is not FDA approved for pIBD at this time.

Etanercept is a dimeric fusion protein consisting of the extracellular ligand-binding portion of the TNF receptor linked to the Fc portion of the human IgG1. It was first approved for the treatment of RA in 1998 and for treatment of JIA 1 year later. Although etanercept has several approved indications in adults with inflammatory arthritis, it is ineffective for the treatment of IBD.

It is unknown how many unique pediatric patients have thus far been exposed to these 3 biologics. According to the FDA, approximately 9200 children and adolescents aged 0–17 years received etanercept through 2007 (source: Amgen), and 2636 patients aged 0–16 years received adalimumab for the 2-year period of 2006–2007 [9]. Approximately 19 000 patients aged 0–18 have been exposed to infliximab between 1998 and 2010 worldwide (source: Centocor). Despite the use of biologics by thousands of children and its important role in the management of patients with JIA and pIBD, the concerns of infectious risk with TNF-α inhibitor therapy continue to be largely extrapolated from adult studies.

METHODS

We performed a systematic literature review to find original studies that evaluated TNF-α inhibitor therapy (infliximab, etanercept, or adalimumab) in pediatric JIA or pIBD patients (Figure 1). We reviewed all references found through PubMed (http://www.ncbi.nlm.nih.gov/pubmed) using a combination of the search terms JIA, CD, UC, IBD, infliximab, etanercept, and adalimumab. The search was limited to studies including human subjects, subjects aged 0–18 years, all clinical trials, case reports, and the English language that were published between 2000 and 2012. We then checked the references of each study to find any missed studies. Studies that did not mention collecting data on the occurrence of infection or that did not describe any infection in the manuscript were excluded. Case reports, case series, and FDA reports were included in the microbiology tables only in order to understand the breadth of serious infections as they are rare events that may be missed in the small pediatric studies available. We defined serious infections as life threatening, requiring hospitalization, requiring intravenous antimicrobial therapy, and/or characterized in individual studies as a serious infection or serious adverse event. Mild infections were defined as infections that did not require hospitalization and/or intravenous antimicrobial therapy. If infection data were unclear, an email was sent to the corresponding author.

Figure 1.

Flow diagram of literature search and selection process. Abbreviations: FDA, Food and Drug Administration; JIA, juvenile idiopathic arthritis; pIBD, pediatric inflammatory bowel disease; TNF-α, tumor necrosis factor alpha.

JUVENILE IDIOPATHIC ARTHRITIS

JIA is the most common chronic childhood rheumatic disease, affecting nearly 300 000 children in the United States. It is a heterogeneous autoimmune disease with 7 categories, which are based on the number of joints affected during the first 6 months of disease and extra-articular involvement. JIA is a diagnosis of exclusion, characterized by arthritis that persists for at least 6 weeks and occurs before age 16. The categories of oligoarticular, polyarticular rheumatoid factor negative, and polyarticular rheumatoid factor positive JIA comprise 65%–80% of all JIA patients [10]. The remaining 4 categories include systemic-onset JIA, enthesitis-related arthritis, psoriatic arthritis, and undifferentiated arthritis.

JIA is a multifactorial disease with many patients likely having a genetic predisposition, but with disease often developing after various environmental triggers that are not yet fully understood [11]. Although still labeled as a JIA subcategory at this time, systemic-onset JIA likely results from a distinct pathogenesis in comparison to all other subtypes [12]. Not surprisingly, treatment is also different for systemic-onset JIA, and TNF-α inhibitors are less efficacious in this cohort [13]. JIA patients of the remaining subtypes are thought to have an autoreactive immune response triggered by the adaptive immune system toward a self-antigen. This includes an uncontrolled response by Th1 and Th17 cells, leading to proliferation of T cells and proinflammatory cytokines with subsequent joint destruction. Recent studies have shown that biologics may lead to clinical response in rheumatologic patients by restoring the balance between specific T-cell populations [14–16]. However, these immune effects may in turn lead to an increased risk of infections from mycobacteria, intracellular bacteria, and fungi [17, 18].

EPIDEMIOLOGY OF INFECTIONS IN JIA PATIENTS TREATED WITH TNF-α INHIBITORS

Frequency and Sites of Mild and Severe Infections

Five case reports, 2 case series, 1 FDA report, 19 prospective studies, and 6 retrospective studies were reviewed (Table 2) [18–50]. Mild infections occurred more frequently and were observed in 8% (2/25) to 97% (31/32) of JIA patients treated with TNF-α inhibitors [31, 36]. Upper respiratory tract infections were most often reported (Table 3). Severe infections occurred in 0% (0/25) to 9% (3/32) of pediatric patients [31, 41]. The most commonly reported sites of severe infections were the respiratory tract and musculoskeletal system (Table 4). The incidence of severe infections in adult RA patients treated with biologics is similar at 3.8%–6.2%, with the organ systems most commonly affected being the respiratory tract and skin [2, 51–54].

Table 2.

Studies That Reported Infections in Juvenile Idiopathic Arthritis Patients Treated With Tumor Necrosis Factor–α Inhibitors

Study, First Author	TNF-α Inhibitor	No. of Subjects	Age, y, Mean	Disease Duration, y, Mean	DMARDa, CS, or Both	Patient-years of Drug Exposure	Total No. of Infections (per Patient-year)	Total No. of Serious Infections (per Patient-year)
Prospective studies
Lovell [34]	ET	69	10.5	5.9	CS	0.54	NS	1 (1.85)
Lovell [32]	ET	58	10.5	5.9	Both	53	78 (1.5)	8 (0.15)
Lovell [45]	ET	26	10.5	5.9	Both	126	NS	10 (0.03)
Rupertob [23]	INX	117	11	4.2	DMARD	99	78 (0.78)	7 (0.07)
Tzaribachev [41]	ET	25	12	5	Both	39.6	2 (0.05)	0
Gerloni [25]	INX	68	21.7	13.7	Both	126	7 (0.06)	1 (0.008)
	ET	95	13.7	8.4		190	34 (0.18)	2 (0.011)
Lovell [33]	AD	86	11.1	3.6	CS	93	97 (0.66)	4 (0.027)
	AD + MTX	85	11.4	4	CS	109	112 (0.65)	3 (0.017)
Giannini [26]	ET	103	10.8	4.8	Both	224	NS	4 (0.018)
	MTX	197	9.0	1.7		388	NS	5 (0.013)
	ET + MTX	294	10.1	3.3		635	NS	13 (0.02)
Horneff [28]	ET	100	13.8	5.5	Both	NS	34	25
	ET + MTX	504	12.5	4.9		NS	50	1
Horneff [29]	ET	20	12.9	4.1	Both	4.6	13 (2.82)	0
Prince [30]	ET	146	11c	4.1c	Both	312	21 (0.07)	4 (0.013)
Ruperto [46]	INX	78	NS	NS	DMARD	171	57 (0.33)	2 (0.012)
Zuber [47]	ET	188	19	4.3	Both	393	959 (2.44)	16 (0.04)
Trachana [35]	AD	26	12.6	7.1	Both	72	10 (0.14)	3 (0.04)
Otten [27]	ET	262	12.4c	3.0	Both	323	99 (0.31)	7 (0.03)
Tynjalab [24]	INX	19	10.5	0.13	DMARD	19.7	36 (1.82)	0
	COMBOd	20	8.3	0.19		20	35 (1.75)	0
	MTX	20	10.1	0.15		20	48 (2.4)	3 (0.15)
Wallaceb,e [48]	ET + MTX	42	9.9	0.41	Both	24.8	16 (0.65)	NS
	MTX	43	11.1	0.43		20.6	18 (0.87)	NS
Mori [31]	ET	19	13.7	6.1	Both	118	99 (0.84)	3 (0.03)
Imagawab [49]	AD	25	13	4.7	Both	29	21 (0.72)	3 (0.1)
Retrospective studies
Takei [38]	ET	8	8.4	5.3	DMARD	10.3	3 (0.29)	0
Tynjala [39]	AD	20	13.4	10	DMARD	31	30 (0.97)	0
Southwoodb [40]	ET	483	12	5	DMARD	941	5 (0.005)	5 (0.005)
Lamot [37]	INX, ET	41	6.9	4.1	DMARD	106	NS	2 (0.02)
Beukelman [50]	ET, INX, or AD	1315	NS	9.7	NS	1580	NS	55 (0.04)
Bracaglia [36]	ET	25	3.3c	1.2c	Both	NS	NS	2

Abbreviations: AD, adalimumab; CS, corticosteroid; DMARD, disease-modifying antirheumatic drug; ET, etanercept; INX, infliximab; MTX, methotrexate; NS, not specified; TNF-α, tumor necrosis factor alpha.

^a DMARD may include 5-aminosalicylic acid, 6-mercaptopurine, cyclosporine, azathioprine, or methotrexate.

^b Excluded subjects previously exposed to any TNF-α inhibitor.

^c Median.

^d Patients on combination of DMARDs.

^e Only study to include 2 or fewer JIA categories. All other studies included 3 or more JIA categories.

Table 3.

Mild Infections in Juvenile Idiopathic Arthritis Patients Treated With Tumor Necrosis Factor–α Inhibitors

Infection Type (No. of Infections)	Infliximab (n = 296)a	Etanercept (n = 2465)	Adalimumab (n = 242)
Respiratory tract infections (1255)
URTIb (1251)	73	1016c	162
LRTI (4)	0	4	0
Skin and soft tissue infections (43)	0	39	4
Gastrointestinal infections (59)	3	56d	0
Genitourinary infections (28)	0	28	0
Viral infections (197)
Primary varicella (13)	0	12	1
Zoster (15)	1	12	2
HSV (49)	1	47	1
EBV (4)	0	4	0
Nonspecific (116)	0	0	116
Fungal infectionse (8)	2	5	1
Mycobacterial infections (1)
Pulmonary tuberculosis	1	0	0
Fever of unknown source (28)	1	27	0

Abbreviations: EBV, Epstein-Barr virus; HSV, herpes simplex virus; LRTI, lower respiratory tract infection; URTI, upper respiratory tract infection.

^a Total number of subjects included in the studies where microbiology and/or infection type was available.

^b Includes pharyngitis, otitis, and sinusitis.

^c Seven cases of influenza.

^d One case of hepatitis A.

^e Includes oropharyngeal and vulvovaginal candidiasis.

Table 4.

Severe Infections in Juvenile Idiopathic Arthritis Patients Treated With Tumor Necrosis Factor–α Inhibitors

Infection Type (No. of Infections)	Infliximab (n = 296)	Etanercept (n = 2465)	Adalimumab (n = 242)
Respiratory tract infections (37)
URTI (14)	0	12	2
LRTI (23)	5a	15	3
Sepsis/bacteremia (8)	0	7b	1c
CNS infections (4)
Meningitis (4)	0	4d	0
Gastrointestinal infections (11)
Gastroenteritis/colitis (11)	0	11e	0
Musculoskeletal infections (25)
Necrotizing fasciitis (2)	0	2f	0
Abscess/cellulitis (15)	1	13	2h
Pyomyositis (2)	1	1	0
Septic arthritis (3)	0	3i	0
Osteomyelitis (2)	1i	1	0
Osteomyelitis and septic arthritis (1)	0	1i	0
Genitourinary infections (11)	1	10	0
Viral infections (17)
Primary varicella (2)	0	2	0
Zoster (9)	2	5	2
HSV (2)	1	0	1
CMV (2)	1	1	0
EBV (1)	0	1	0
HBV (1)	0	0	1j
Fungal infections (4)
Histoplasmosis (4)	2	2	0
Mycobacterial infections (4)
Pulmonary tuberculosis (1)	1	0	0
Extrapulmonary tuberculosis (3)	1	2	0
Infectious fatalities (4)	1	2	1

Abbreviations: CMV, cytomegalovirus; CNS, central nervous system; EBV, Epstein-Barr virus; HBV, hepatitis B virus; HSV, herpes simplex virus; LRTI, lower respiratory tract infection; URTI, upper respiratory tract infection.

^a One case of Streptococcus pneumoniae.

^b Two cases of group A Streptococcus (GAS) purpura fulminans, both fatal.

^c Fatal infection.

^d One case each of varicella-zoster virus (VZV) and EBV meningoencephalitis.

^e One case each of Escherichia coli and Clostridium difficile colitis.

^f One case of Staphylococcus aureus infection; 1 case secondary to VZV infection.

^g One case of Enterococcus faecalis urachal cyst infection.

^h Both cases due to S. aureus infection in same patient.

ⁱ One case of GAS infection.

^j Primary HBV infection. Patient was negative for HBV infection prior to starting adalimumab.

Overall, the rates of mild and severe infections observed in JIA patients treated with biologics appear significant, but wide-ranging, especially for mild infections. The incidence of infections likely varied in the studies reviewed due to the inclusion of a heterogeneous JIA population with varying subtypes of JIA, and disease duration, small numbers of patients enrolled, use of concurrent DMARDs and/or corticosteroids, and limitations of study design that focused more on efficacy (Table 2). Importantly, it is unknown whether JIA itself further contributes to an increased risk of infection. A recent study found that there may be an increased rate of hospitalization with bacterial infections in JIA patients compared to healthy children [50]. These authors also found no increased rate of hospitalized bacterial infections among JIA patients treated with TNF-α inhibitors.

Microbiology

The microbiology of infections was unavailable in most of the pediatric studies reviewed. Bacterial pathogens most often identified were Streptococcus pyogenes and Staphylococcus aureus. Herpes simplex virus and varicella zoster virus (VZV) were the most frequently identified viral infections in both mild and severe illnesses. Three children who developed severe VZV disease (aseptic meningitis, disseminated varicella, and necrotizing fasciitis), were nonimmune prior to starting anti-TNF therapy [32, 36].

Fungal and mycobacterial infections were also reported. Progressive disseminated histoplasmosis was described in 4 patients and all lived in an endemic region [42]. Only 1 patient was on concurrent corticosteroids. Five cases of M. tuberculosis were reported [18, 23, 43]. These opportunistic infections have similarly been reported in adults [3, 42, 55]. Other opportunistic infections described in RA patients treated with TNF-α inhibitors include listeriosis, aspergillosis, and Pneumocystis jirovecii pneumonia, which have not yet been described in JIA patients [55, 56]. Differences in microbiology observed in RA compared to JIA patients is likely multifactorial, including longer disease duration, older age, higher cumulative exposure to immunosuppressive therapy, and the underlying disease [57].

Pediatric IBD

There are approximately 150 000 pIBD patients aged 0–17 years in the United States [58]. UC and CD are inflammatory bowel diseases characterized by inflammation of the gastrointestinal tract. UC involves recurring inflammation of the mucosal layer of the colon, almost invariably involving the rectum, but may affect any portion of the colon in a continuous fashion. CD is characterized by transmural inflammation of any component of the gastrointestinal tract from the oral cavity to the anus. Although UC and CD have distinct pathologic and clinical characteristics, both emerge from genetic and environmental influences that likely stem from an abnormality in mucosal immune function [59].

In comparison with adult-onset disease, pediatric UC patients tend to have more extensive intestinal involvement and a more severe disease course, and are more likely to be corticosteroid dependent [60]. This may also be true of pediatric CD, but has not been consistently shown [60, 61]. In pIBD, both the disease and corticosteroid therapy may result in significant long-term adverse effects such as poor weight gain and linear growth. Thus, treatment with TNF-α inhibitors has significantly altered the outcomes for pIBD patients by allowing for steroid-sparing regimens and improved linear growth [62].

Infliximab, adalimumab, and certolizumab have been shown to be efficacious in treating IBD. Because no studies evaluating certolizumab therapy in pIBD could be found at the time of this review, it is not included. Although recent studies have shown infliximab and adalimumab to be effective in inducing and maintaining remission in pIBD, only infliximab is currently FDA approved for pIBD (CD and UC) [63]. Nevertheless, adalimumab and certolizumab have been used off-label to treat pIBD.

EPIDEMIOLOGY OF INFECTIONS IN pIBD PATIENTS TREATED WITH TNF-α INHIBITORS

Frequency and Sites of Mild and Severe Infections

Eight case reports, 1 case series, 2 FDA reports, 11 prospective studies, and 17 retrospective studies describing the incidence of infections in pIBD patients treated with infliximab and/or adalimumab were reviewed (Table 5) [63–101]. Most infectious adverse events reported in pIBD patients were mild infections, specifically upper respiratory tract infections (Table 6). Sepsis, gastrointestinal, and soft tissue infections were the most common types of severe infections (Table 7).

Table 5.

Studies That Reported Infections in Pediatric Inflammatory Bowel Disease Patients Treated With TNF-α Inhibitors

Study, First Author	IBD Type	TNF-α Inhibitor	No. of Subjects	Age, y, Mean	Disease Duration, y, Mean	DMARDa, CS, or Both	Patient-years of Drug Exposureb	Total No. of Infections (per Patient-year)	Total No. of Serious Infections (per Patient-year)
Prospective
Kugathasan [77]	CD	INX	15	12.8	10.8	Both	NS	0	0
Baldassano [78]	CD	INX	21	15.0c	3.8c	Both	NS	3	0
Cezard [79]	CD	INX	21	15.0c	4.0	Both	NS	1	1
Hyams [81]	CD	INX	112	13.3	2.0	Both	85	70 (0.824)	9 (0.106)
Hyams [72]	CD	INX	202	12.7	NS	Both	NS	1	1
Ruemmele [83]	CD	INX	40	13.9	3.0	Both	NS	15	0
Viola [89]	CD	ADA	23	16.1c	4.4	Both	NS	8	1
Hyams [66]	UC	INX	52	13.3	NS	Both	NS	0	0
Crombe [71]	CD	INX	120	14.5c	9.3c	NS	320d	2 (0.006d)	0
Hyams [63]	CD	INX	60	13.2	1.8	Both	98	54 (0.551)	8 (0.082)
Hyams [85]	UC	INX	60	14.5c	1.4c	Both	34	33 (0.971)	6 (0.176)
Retrospective
Serrano [100]	Both	INX	18	15.5	4.9	Both	12d	2 (0.168d)	2 (0.168d)
Stephens [64]	CD	INX	82	15.2	NS	Both	73d	3 (0.041d)	0
De Ridder [76]	CD	INX	30	14.1	NS	Both	63d	1 (0.016d)	1 (0.016d)
Friesen [65]	Both	INX	111	13.4	NS	Both	176d	4 (0.023d)	0
Lamireau [80]	CD	INX	88	11.0	1.2	NS	29	2 (0.069)	2 (0.069)
Mamula [67]	UC	INX	8	15.3	1.5c	Both	6	1 (0.159)	0
Eidelwein [68]	UC	INX	12	12.5c	1.6c	Both	NS	1	1
Wewer [75]	CD	INX	24	15.4c	2.2c	Both	NS	1	0
Cucchiara [69]	UC	INX	22	12c (M); 14c (F)	NS	Both	NS	0	0
De Ridder [74]	CD	INX	66	14.5	2.4	Both	159	10 (0.063)	0
McGinnis and Murray [70]	UC	INX	40	12.9	0.54c	Both	62d	3 (0.049d)	1 (0.016d)
Wynands [82]	CD	INX	38	10.7	3.1	NS	22	1 (0.045)	0
Wyneski [99]	CD	ADA	15	16.6	5.7	Both	12	4 (0.333)	0
Rosh [88]	CD	ADA	115	11.1	NS	Both	127d	2 (0.016d)	1 (0.008d)
Rosenbach [87]	CD	ADA	14	13.9c	3.9c	NS	17	1 (0.059)	1 (0.059)
De Bie [73]	CD	INX	152	15.0c	1.8c	Both	203	27 (0.133)	2 (0.010)
Russell [86]	Both	ADA	72	14.8c	4.3c	NS	NS	5	4

Abbreviations: ADA, adalimumab; CD, Crohn's disease; CS, corticosteroids; DMARD, disease-modifying antirheumatic drug; F, female; IBD, inflammatory bowel disease; INX, infliximab; M, male; NS, not specified; TNF-α; tumor necrosis factor alpha; UC, ulcerative colitis.

^a DMARD may include 5-aminosalicylic acid, 6-mercaptopurine, cyclosporine, azathioprine, or methotrexate.

^b Calculated based on patient-years of exposure to TNF-α inhibitor.

^c Median years.

^d Calculated based on patient-years of follow-up in study. Unable to determine patient-years of exposure to TNF-α inhibitor from study.

Table 6.

Mild Infections in Pediatric Inflammatory Bowel Disease Patients Treated With Tumor Necrosis Factor–α Inhibitors

Infection Type (No. of Infections)	Infliximab (n = 1407)	Adalimumab (n = 241)
Respiratory tract infections (114)
URTI (105)	94	11
LRTI (9)	9	0
Skin and soft tissue infections (14)	13a	1b
Gastrointestinal infections (4)	4	0
Other viral infections (15)
Primary varicella (1)	1	0
Zoster (12)	12	0
HSV (1)	1	0
EBV (1)	1	0
Mycobacterial infections (1)	0	1c
Mild infections not specified (71)	71	NR

Abbreviations: EBV, Epstein-Barr virus; HSV, herpes simplex virus; LRTI, lower respiratory tract infection; NR, not reported; URTI, upper respiratory tract infection.

^a Three cases of dermatophytoses.

^b One case of Staphylococcus aureus.

^c One case of Mycobacterium avium complex.

Table 7.

Severe Infections in Pediatric Inflammatory Bowel Disease Patients Treated With Tumor Necrosis Factor–α Inhibitors

Infection Type (No. of Infections)	Infliximab (n = 1407)	Adalimumab (n = 241)
Respiratory tract infections (4)
URTI (1)	1	0
LRTI (3)	3	0
Sepsis/bacteremia (10)
Sepsis (9)	7a	2b
Bacteremia (1)	1c
CNS infections (2)
Meningitis (2)	1d	1d
Gastrointestinal infections (9)
Abdominal abscess (6)	3	3
Colitis (3)	2	1e
Musculoskeletal infections (9)
Abscess/cellulitis (8)	8c,f	0
Septic arthritis and osteomyelitis (1)	1c	0
Genitourinary infections (1)	1	0
Primary varicella (3)	3g	0
CMV (1)	1h	0
EBV (1)	1	0
Fungal infections (9)
Histoplasmosis (6)	4	2
Aspergillosis (1)	0	1i
PCP (1)	1j	0
Candidemia (1)	0	1k
Infectious fatalities (4)	2h	2b

Abbreviations: CMV, cytomegalovirus; CNS, central nervous system; EBV, Epstein-Barr virus; LRTI, lower respiratory tract infection; PCP, Pneumocystis jirovecii pneumonia; URTI, upper respiratory tract infection.

^a One case each of Escherichia coli and Staphylococcus aureus sepsis.

^b Bacteremia associated with central lines; 1 case due to E. coli and 1 case due to coagulase-negative Staphylococcus infection. Both were fatal infections.

^c One case each of S. aureus infection.

^d One case each of Listeria monocytogenes.

^e One case of Clostridium difficile colitis.

^f One case of Enterococcus faecalis infection.

^g Disseminated varicella-zoster virus infection in 2 patients.

^h Disseminated CMV infection that was fatal.

^I Had concurrent coagulase-negative Staphylococcus line sepsis.

^j Concurrent PCP and histoplasmosis in same patient.

^k Associated with central line infection.

In pIBD patients treated with either adalimumab or infliximab, the incidence of mild infections ranged from 3% (1/38) to 77% (46/60), and from 0% (0/66) to 10% (6/60) for serious infections [63, 65, 74]. In adult IBD patients, the risk of serious infections with TNF-α inhibitor therapy is between 2.2% to 5% [102–104]. However, the adult IBD studies have had conflicting results because these patients may also be at increased risk of serious infections due to the underlying disease and advanced age [105]. It is unknown if pIBD patients have a similar increased risk of infections because of their underlying illness.

Microbiology

The microbiologic results of mild and severe infections in pIBD patients treated with TNF-α inhibitors were extremely limited; most studies had no information on mild infections or pathogens involved. For both mild and severe infections, VZV was the most commonly identified viral pathogen. There were 2 reports of disseminated varicella and both patients either had evidence of immunity or prior exposure [84].

Bacterial and fungal infections also contributed to severe infections in pIBD patients. There were 2 cases of Listeria monocytogenes meningitis in 1 CD patient and 1 UC patient treated with infliximab and concurrent DMARDs [92, 101]. Histoplasmosis was described in 6 patients, all of whom lived in an endemic region and 5 of whom were on concurrent DMARDs [94–96]. One patient was on infliximab only and presented concurrently with P. jirovecii pneumonia [95]. A recent study in the murine model found that TNF-α inhibitors may alter protective immunity and dampen the response to Histoplasma capsulatum infection [106]. There was 1 report of systemic Mycobacterium avium complex infection [93].

Two infectious fatalities from infliximab were reported in CD patients: 1 due to disseminated cytomegalovirus and 1 due to bacterial sepsis [73, 97]. There were 2 infectious fatalities with adalimumab [86]. Both patients required central venous catheters for parenteral nutrition and were on other DMARDs. One patient had coagulase-negative staphylococcal bacteremia followed by pulmonary invasive aspergillosis.

Approaches to Screening and Prevention of Infections in Children With JIA or pIBD Treated With TNF-α Inhibitors

Recommendations for screening strategies, vaccinations, and safety monitoring of pediatric patients on TNF-α antagonists are scarce and largely extrapolated from adult studies. The American College of Rheumatology recommends that JIA patients be screened for latent tuberculosis, and if high risk, chronic hepatitis B (HBV) and hepatitis C (HCV) infection, before starting biologics [13]. Similarly, pIBD patients should be screened for latent tuberculosis prior to initiating TNF-α inhibitor therapy, and then annually. This has likely helped minimize the incidence of these opportunistic infections in pediatric patients, with only 5 cases of M. tuberculosis and no reports of HBV or HCV reactivation found in this cohort.

Vaccine recommendations are inadequately detailed for pIBD and JIA patients who are starting TNF-α inhibitors. For adults with RA or IBD, there are more specific preventive measures, including the recommendation to provide influenza vaccine annually, as well as pneumococcal, HBV, and human papillomavirus (HPV) vaccines [107, 108]. In pIBD patients, the European Crohn's and Colitis Organization recommends administering routine inactivated vaccines, including HBV and HPV [109]. In pediatrics, only 1 of 3 JIA guidelines found mentioned the importance of updating the patient on vaccines prior to initiation of TNF-α inhibitors [13, 110, 111]. Although there may be a concern for flares in JIA patients with vaccine administration, studies have shown that JIA patients who received hepatitis A, HBV, meningococcal C, or measles-mumps-rubella (MMR) vaccines were not at increased risk of reactivation of their underlying disease [112–115]. Ideally, children with pIBD and JIA should adhere to the immunization schedule recommended by the American Academy of Pediatrics, except for the administration of live vaccines, which are currently contraindicated while on biologics. JIA and pIBD patients should be caught up with vaccines and complete the regimen of live attenuated vaccines prior to immunosuppressive therapy whenever possible.

In our practice, we evaluate the patient's varicella and measles immune status prior to initiation of TNF-α inhibitor therapy. We encourage VZV vaccination if a patient is nonimmune prior to TNF-α inhibitor therapy, given the risk of severe varicella disease and zoster as described above. If a patient is MMR nonimmune and he/she has only received 1 dose, we recommend administering the second MMR dose at least 4 weeks prior to initiating TNF-α inhibitor therapy. This is especially important given recent measles outbreaks in the United States and possible diminished immunogenicity to vaccines while on biologics [116].

Because of the incidence of histoplasmosis observed with TNF-α inhibitor therapy, all patients should be screened for risk factors prior to starting therapy, and then annually. This includes questioning patients about travel to endemic regions and participation in activities that may increase risk of exposure. Although there are no formal recommendations on screening strategies in high-risk patients, these patients may warrant closer monitoring.

CONCLUSIONS

Pediatric patients with JIA and IBD can frequently develop mild infections and, less commonly, severe infections when treated with infliximab, etanercept, or adalimumab. Bacterial, viral, and fungal infections were all important etiologies of serious infections. Importantly, few pediatric patients developed M. tuberculosis, likely owing to effective screening for latent tuberculosis. Unfortunately, the majority of the studies reviewed had significant limitations, making it difficult to adequately assess the extent of infections that occurred in this cohort. This includes limited information about the frequency, sites of infection, and microbiology. As mentioned, studies included heterogeneous populations on concomitant immunosuppressive therapy and were not designed to effectively evaluate for mild and/or severe infections. Future prospective studies with larger patient populations, more frequent follow-up, and a more thorough assessment of mild and serious infections would significantly help clinicians better understand the implications of starting a patient with JIA or pIBD on TNF-α inhibitors, and ultimately help to improve prevention strategies and management.