|
Indian Pediatr 2015;52: 883-888 |
|
Is Antibiotic Exposure Associated With Newly
Diagnosed Juvenile Idiopathic Arthritis?
|
Source Citation: Horton DB, Scott FI, Haynes K, Putt
ME, Rose CD, Lewis JD, et al. Antibiotic exposure and juvenile
idiopathic arthritis: A case-control study. Pediatrics.
2015;136:e333-43.
Section Editor: Abhijeet Saha
|
Summary
In this nested case-control study in a
population-representative medical records database from the United
Kingdom, children with newly diagnosed juvenile idiopathic arthritis
(JIA) were compared with age- and gender-matched control subjects
randomly selected from general practices containing at least one case,
excluding those with inflammatory bowel disease, immuno-deficiency, or
other systemic rheumatic diseases. Conditional logistic regression was
used to examine the association between antibacterial antibiotics
(including number of antibiotic courses and timing) and JIA after
adjusting for significant confounders. Any antibiotic exposure was
associated with an increased rate of developing JIA [adjusted OR 2.1
(95% CI 1.2, 3.5)]. This relationship was dose dependent [adjusted OR
over 5 antibiotic courses 3.0 (95% CI 1.6,5.6)], strongest for exposures
within 1 year of diagnosis, and did not substantively change when
adjusting for number or type of infections. In addition,
antibiotic-treated upper respiratory tract infections were more strongly
associated with JIA than untreated upper respiratory tract infections.
The authors concluded that antibiotics were associated with newly
diagnosed JIA in a dose- and time-dependent fashion. Antibiotic exposure
may play a role in JIA pathogenesis, perhaps mediated through
alterations in the microbiome.
Commentaries
Evidence-based Medicine Viewpoint
Relevance: Juvenile idiopathic arthritis (JIA) is
globally recognized as the commonest chronic musculoskeletal disease
affecting children [1]. The individual and societal impacts of this
chronic condition on quality of life, productivity and health-care costs
have been well documented [1,2]. Although it is well recognized as an
autoimmune condition, its precise etiology and the complex interplay of
factors causing and/or worsening the clinical condition are unclear. A
certain degree of similarity is anticipated in the causal pathway of
diverse conditions that broadly fit under the umbrella of autoimmune
diseases. It is therefore not surprising that perturbations though to be
involved in inflammatory bowel disease, type I diabetes mellitus, celiac
disease would be explored in rheumatic conditions as well. This recent
publication [3] reported a case-control study exploring the relationship
between usage of antibiotics in early childhood and the subsequent
diagnosis of JIA.
Critical appraisal: The case-control study design
is used to study association between exposure to ‘risk factors’ and the
occurrence of disease. Briefly, it involves recruitment of people with
the outcome of interest (disease) designated as cases, and those without
the outcome (disease) designated as controls; and working backwards in
time to identify their exposure status. This design is especially useful
to study relatively rare outcomes (diseases) particularly if the time
period between exposures to outcome is long. Although it is placed
relatively low on the evidence hierarchy [4] on account of high
susceptibility to bias, it is an excellent design to study the effect of
risk factors to which research participants cannot be ethically exposed.
There are several tools available for critical appraisal of case-control
studies, but in general they revolve around step-wise evaluation of
three issues viz appraisal of validity, assessment of results and
exploration of local applicability. Appraisal using such tools is
summarized in Tables I and II [5,6].
TABLE I Critical Appraisal of the Study
Criteria |
Report |
Did
the study address a clearly focused issue? |
Yes.
The investigators clearly specified their objective to examine
the relationship between usage of antibiotics (E=Exposure), and
subsequent confirmation of JIA (O=Outcome) in children
(P=Population), compared to a matched population without
exposure to antibiotics (C=Comparison). |
Did
the authors use an appropriate method to answer their question?
|
Yes.
A case-control study is an acceptable way to address the
research question. A comparative cohort study (comparing groups
of children with and without antibiotic usage, for the
development of JIA) would be superior in terms of study design
(as it would allow controlling for various confounding factors
thereby limiting bias), but too complex and cumbersome in terms
of resources, logistics and time. |
Were
the cases recruited in an acceptable way? |
Yes
Cases were identified from a population medical record database
in the UK covering about 550 general practices. In the
database, JIA was defined with a code analogous to its ICD 10
code. The age range of the population of interest was 1-15
years. Efforts were made to increase the specificity of the
system by examining the database for alternate definitions of
the outcome of interest adding the following (singly or in
combination) to JIA: prescription of NSAID or steroids during 2
months prior to one year after the diagnosis, use of any
disease-modifying agents, referral to specialist
rheumatology services. There were limited exclusion criteria,
thereby minimizing selection bias. It is presumed that all
children with JIA would be picked up through the electronic
database; hence completeness of case identification is not a
concern. Similarly, the system is expected to include cases of
all severities hence selection bias by severity is also
addressed effectively. Since data were retrieved through
records, the risk of missing cases is limited unless cases with
JIA were mislabeled as other disorders. Of course it is unclear
if children with currently (or subsequently) diagnosed JIA could
actually have other conditions mislabeled as JIA.
|
Were
the controls recruited in an acceptable way? |
Yes
The investigators recruited 10 age- and gender-matched children
without JIA (as described above) for each case. The controls
were recruited from the same general practices as cases. As
controls were identified using the electronic database and not
physically recruited, bias due to non-response is absent. There
is no well-defined optimal ratio of cases to controls, but in
general, statistical power improves with more controls although
the effect plateaus at about 4 controls per case. |
Was
the exposure accurately measured to minimize bias? |
Yes.
The electronic database was used to retrieve information about
antibiotic prescriptions, including data on antibiotic name,
dosage and duration. Presumably there is no other way (than
prescriptions) for children to obtain antibiotics in the
population studied. The electronic database precludes
measurement and recall biases to exposure. The same method to
assess exposure was used for cases and controls. |
|
|
What
confounding factors have the authors accounted for? |
The
investigators attempted to identify and take into account
several potential confounding factors including demographic
features, frequency and type of preceding infections, prevalence
of auto-immune disease, maternal auto-immune disease, and
frequency of hospitalization. However, the frequency and
duration of breastfeeding as well as ethnicity of recruited
children have not been considered. Further, personal or family
history of immune-deficiency has also not been considered. The
authors did include an index representing (socio-economic)
deprivation, but did not describe results by this variable. Data
analysis included sensitivity analysis as well as regression.
|
What
are the results of this study?
How precise are the results? |
In
summary, the study reported that JIA cases were more likely to
be exposed to antibiotic therapy, greater number of antibiotics,
more antibiotics with entero-hepatic circulation, more
infections, hospitalization, auto-immune diseases, as well as
maternal auto-immune disease. The following adjusted odds ratios
(95% Confidence intervals) were presented: Any antibiotic
prescription: 2.1 (1.2, 3.5); Any infection: 1.8 (0.4, 3.4);
Auto-immune disease in the child: 30.6 (3.4, 278.0). The primary
outcome result was robust even when four alternate case
definitions were used. There was increased risk of developing
JIA with greater number of antibiotic courses, higher total
duration of antibiotic use, and more recent usage of
antibiotics. Most of these results were robust when explored
with alternate case definitions. |
Do
you believe the results? |
The
results are fairly convincing in terms of clinically significant
effect, demonstration of dose dependency, and robustness with
changes in case definition. The study demonstrates the
fulfilment of several of the Bradford Hill criteria for
causation (Table II). |
Can
the results be applied to the local population?
|
Theoretically the results can be applied to any population
unless strong reasons for the contrary can be unequivocally
demonstrated. There are no obvious biological, genetic, social,
cultural, or economic reasons to believe that if antibiotic
exposure is somehow a part of the causal pathway of JIA in this
UK based study, the results would be different in India.
However, it is difficult to quantify the increased risk in the
local population in the absence of baseline prevalence data. |
Do
the results of this study fit with other available evidence? |
Unfortunately, there is very limited data exploring the
relationship between antibiotic exposure and JIA. However, one
recent case-control study [7] examining a country-wide database
over 10 years in Finland with 1298 cases and 5179 age, gender
and residence matched controls; identified odds ratio of
developing JIA with antibiotic exposure to be 1.6 (95% CI 1.3,
1.9). They also reported that early antibiotic exposure (i.e
prior to 2 years of age) increased the risk significantly. There
do not appear to be animal experiments testing antibiotic
exposure and the subsequent development of JIA type symptoms, or
histopathology. |
The investigators used several methodological
refinements to minimize bias and provide robust results. They also tried
to distinguish the effect of antibiotics from the effect of infections,
which was missing in the previous study [7]. They considered several
sources of bias and confounding; and attempted to address them
effectively. For these reasons, the data from this study are considered
reliable. However, inexplicably Table I shows a total of
1672 participants (152 cases + 1520 controls) at one place but only 1662
(1280 exposed + 382 unexposed) at another place. The investigators did
not consider subgroup analyses base on different types of JIA.
Overall this well-conducted study suggests that
antibiotic usage in childhood is associated with a greater risk of
developing JIA in later childhood. In contrast, there are three
well-known pieces of information that complicate the puzzle. First, if
antibiotics are responsible for causation of JIA, it follows that
health-care systems where childhood antibiotic use is rampant (India is
a classic example) should witness the highest incidence/prevalence of
JIA. Unfortunately, there is no robust population-based prevalence data
in India to support or refute this. However, there is a single study [8]
evaluating prevalence of musculoskeletal symptoms in school children
(6-17 y) where one case of JIA was identified among 2059 children
studied. From this, the authors calculated a population prevalence of
48.5/100,000 school children. A similar study in Oman reported a
prevalence of 20 per 100,000 [9]. In contrast, the prevalence in
developed countries is variably reported to range from 100-400 per
100,000 (A), although more recent estimates suggest 50-60/ 100,000 in
one region of the USA [10], 90/100,000 in Sweden (L), and 33.5/100,000
in this study [3]. Of course, the limited Indian data cannot be reliably
compared with other studies on account of methodological differences;
however clinical experience also does not suggest an unusually high
burden of JIA.
If early [7] and more recent [3] antibiotic exposure
increase the risk of developing JIA, it would be expected that countries
with high burden of neonatal sepsis (such as India) would have very high
burden of JIA, and that too presenting at a relatively younger age.
Further, there is no data from developed or
developing countries to suggest that population subgroups exposed to
greater frequency and/or duration of antibiotics (such as children with
primary or secondary immune deficiencies) have higher prevalence of JIA.
Of course, the reverse has been noted i.e children with JIA (or
other autoimmune conditions) have greater frequency of infections
requiring antibiotics [11]. Some authors have reported co-existing JIA
and primary immune deficiencies [12-15] , but it is unclear which came
first, although there is a report of a small cohort of 25 primary
immune-deficiency patients in Turkey who were observed to later develop
(rather be diagnosed with) diverse auto-immune conditions [16].
One other issue is that most auto-immune diseases
(including JIA) have a female preponderance, whereas antibiotic exposure
is generally gender independent. All these points suggest that while
antibiotic usage could contribute, it is unlikely to be the absolute
cause of development of JIA.
Extendibility: The study setting,
population characteristics (ethnicity, genetic background, environmental
exposures, etc) and health-care system are very different from our
country. However, the conclusion that antibiotic usage could be linked
to JIA appears extendible to India also. Unfortunately, the contrary
pieces of evidence complicate the picture, making it difficult to draw a
firm conclusion. Nevertheless, it goes without saying that unnecessary
antibiotic use must be restricted to the maximum extent possible.
Further, all antibiotic therapy must be well documented in terms of
indication, drugs used, dosage and duration.
Conclusions: This well-designed case-control
study suggests that antibiotic use in early childhood could have a
dose-dependent impact on subsequent development of JIA.
References
1. Kumar S. Need for determining the incidence and
prevalence of JIA in developing countries: the Indian predicament.
Rheumatology. 2010;49:1598-9.
2. Minden K, Niewerth M, Listing J, Möbius D, Thon A,
Ganser G, et al. The economic burden of juvenile idiopathic
arthritis-results from the German paediatric rheumatologic database.
Clin Exp Rheumatol. 2009;27:863-9.
3. Horton DB, Scott FI, Haynes K, Putt ME, Rose CD,
Lewis JD, et al. Antibiotic Exposure and Juvenile Idiopathic
Arthritis: A Case-Control Study. Pediatrics. 2015;136:e333-43.
4. No authors listed. Oxford Centre for
Evidence-based Medicine – Levels of Evidence (March 2009). Available
from:
http://www.cebm.net/oxford-centre-evidence-based-medicine-levels-evidence-march-2009/.
Accessed September 14, 2015.
5. Critical Appraisal Skills Programme (CASP): 11
questions to help you make sense of a case control study. Available
from: http://media.wix.com/ugd/dded87_63fb65dd4e0548
e2bfd0a982295f839e.pdf. Accessed September 14, 2015.
6. No authors listed. The Bradford Hill Criteria.
Available from:
http://www.southalabama.edu/coe/bset/johnson/bonus/Ch11/Causality%20criteria.pdf.
Accessed September 14, 2015.
7. Arvonen M, Virta LJ, Pokka T, Kröger L, Vähäsalo
P. Repeated exposure to antibiotics in infancy: a predisposing factor
for juvenile idiopathic arthritis or a sign of this group’s greater
susceptibility to infections? J Rheumatol. 2015;42:521-6.
8. Abujam B, Mishra R, Aggarwal A. Prevalence of
musculoskeletal complaints and juvenile idiopathic arthritis in children
from a developing country: A school-based study. Int J Rheum Dis.
2014;17:256-60.
9. Abdwani R, Abdalla E, Al Abrawi S, Al-Zakwani I.
Epidemiology of juvenile idiopathic arthritis in Oman. Pediatr Rheumatol
Online J. 2015;13:33.
10. Krause ML, Crowson CS, Michet CJ, Mason T,
Muskardin TW, Matteson EL. Juvenile idiopathic arthritis in Olmsted
County 1960-2013. Arthritis Rheumatol. 2015;Aug 28:doi:
10.1002/art.39323. (Epub ahead of print)
11. Salonen PH, Säilä H, Salonen JH, Vuorela M,
Kautiainen H, Lyytikäinen O, et al. Bloodstream infections
among children with juvenile idiopathic arthritis: a prospective study
from the onset of disease. Clin Exp Rheumatol. 2014;32:979-83.
12. Patiroglu T, Akar HH, Gunduz Z, Sisko S, Ng YY.
X-linked agammaglobulinemia in two siblings with a novel mutation in the
BTK gene who presented with polyarticular juvenile idiopathic arthritis.
Scand J Rheumatol. 2015;44:168-70.
13. Váncsa A, Tóth B, Szekanecz Z. BTK gene mutation
in two non-identical twins with X-linked agammaglobulinemia associated
with polyarticular juvenile idiopathic arthritis. Isr Med Assoc J.
2011;13:579-80.
14. Abolhassani H, Amirkashani D, Parvaneh N,
Mohammadinejad P, Gharib B, Shahinpour S, et al. Autoimmune
phenotype in patients with common variable immunodeficiency. J Investig
Allergol Clin Immunol. 2013;23:323-9.
15. Ludvigsson JF, Neovius M, Hammarström L.
Association between IgA deficiency & other autoimmune conditions: A
population-based matched cohort study. J Clin Immunol. 2014;34:444-51.
16. Patiroglu T, Gungor HE, Unal E. Autoimmune
diseases detected in children with primary immunodeficiency diseases:
Results from a reference centre at middle anatolia. Acta Microbiol
Immunol Hung. 2012;59:343-53.
Joseph L Mathew
Department of Pediatrics,
PGIMER, Chandigarh, India.
Email:
[email protected]
Pediatric Rheumatologist’s Viewpoint
In this paper, Horton, et al. make a
compelling case for implicating antimicrobials with evolution of
juvenile idiopathic arthritis (JIA). A recent study from another center
also showed an association, albeit indirect, with JIA [1]. JIA remains
the commonest rheumatic disease of childhood and is associated with
significant morbidity.
There is no gainsaying the fact that antimicrobials
are frequently misused in children. It is also known that use of
antimicrobials in early childhood significantly alters the human
microbiome and may predispose to long-term and, seemingly unrelated,
consequences. These include, amongst others, inflammatory bowel disease
and obesity [2-4].
In this study, authors suggest that use of
antimicrobials in children may have some role to play in the
pathogenesis of JIA. Their conclusion is based on a nested case-control
study from the United Kingdom conducted on an apparently
‘population-representative medical records database’. The data, however,
need to be interpreted with some caution. JIA is a complex clinical
entity [5,6]. Its etiology remains unknown and the pathogenesis poorly
understood. It must also be understood that JIA is by no means a single
disease entity but a conglomerate of disparate clinical conditions
having chronic arthritis as the common denominator. Further, each of
these conditions may have a different etiology [5,6]. At least two types
of JIA are considered to be rather distinctive in their clinical
presentation and pathogenesis. For instance, it is now well known that
systemic JIA (sJIA) is more of an autoinflammatory rather than an
autoimmune disorder [7]. Similarly the clinical phenotype of enthesitis
related arthritis (ERA) is very different from, say, polyarticular or
oligoarticular JIA. ERA has a distinct association with preceding
gastrointestinal or genito-urinary infections while the other types of
JIA do not [8,9]. Horton, et al. do not specify how many of their
patients had sJIA or ERA. It is difficult to interpret the data in
absence of this information. It may perhaps have been more prudent to
exclude these two categories of JIA from final analysis.
Nevertheless paediatricians, and especially
paediatric rheumatologists, would do well to keep the results of this
study in mind the next time they see a child with JIA, and the next time
they consider prescribing an antimicrobial to a child!
References
1. Arvonen M, Virta LJ, Pokka T, Kröger L, Vähäsalo
P. Repeated exposure to antibiotics in infancy: a predisposing factor
for juvenile idiopathic arthritis or a sign of this group’s greater
susceptibility to infections? J Rheumatol. 2015; 42:521-6.
2. Conte MP, Schippa S, Zamboni I, Penta M, Chiarini
F, Seganti L, et al. Gut-associated bacterial microbiota in
paediatric patients with inflammatory bowel disease. Gut.
2006;55:1760-7.
3. Bailey LC, Forrest CB, Zhang P, Richards TM, Livshits
A, DeRusso PA. Association of antibiotics in infancy with early
childhood obesity. JAMA Pediatr. 2014;168:1063-9.
4. Jernberg C, Löfmark S, Edlund C, Jansson JK.
Long-term impacts of antibiotic exposure on the human intestinal
microbiota. Microbiology. 2010; 156:3216-23.
5. Consolaro A, Varnier GC, Martini A, Ravelli A.
Advances in biomarkers for paediatric rheumatic diseases. Nat Rev
Rheumatol. 2015;11:265-75.
6. Hersh AO, Prahalad S. Immunogenetics of juvenile
idiopathic arthritis: A comprehensive review. J Autoimmun. 2015; Aug
21:doi: 10.1016/j.jaut.2015.08.002. [Epub ahead of print]
7. Martini A. Systemic juvenile idiopathic arthritis.
Autoimmun Rev. 2012; 12:56-9.
8. Ramanathan A, Srinivasalu H, Colbert RA. Update on
juvenile spondyloarthritis. Rheum Dis Clin North Am 2013; 39:767-88.
9. Stoll ML, Kumar R, Morrow CD, Lefkowitz EJ, Cui
X, Genin A, et al. Altered microbiota associated with abnormal
humoral immune responses to commensal organisms in enthesitis related
arthritis. Arthritis Res Therap. 2014;16:486.
Surjit Singh
Department of Pediatrics,
PGIMER, Chandigarh, India.
Email:
[email protected]
Infectious Disease Specialist’s Viewpoint
Investigators have used the Health Improvement
Network, a population-based medical records database in the United
Kingdom that contains comprehensive diagnostic and outpatient
prescription data, to identify people younger than 16 years of age who
were newly diagnosed with arthritis. The 152 children with juvenile
arthritis were matched, for age and sex, with 1520 control subjects from
general practices in the United Kingdom. They looked for antibiotic
prescriptions in these 152 children in the last one year. The risk of
developing arthritis was found to increase as exposure to antibiotics
increased. However, there was no association between the development of
arthritis and exposure to nonbacterial antimicrobial agents, including
antifungal and antiviral drugs. After adjustment for the number and type
of infections children had, the associations did not change
significantly. The age at which children were exposed to antibiotics
also had no significant effect on the associations.
Authors suggest that alterations in the human
microbiome might be implicated in the development of autoimmune
diseaases; and that includes inflammatory bowel disease and rheumatoid
arthritis and perhaps psoriatic arthritis; all of which have some common
features with juvenile arthritis. If the association between antibiotics
and juvenile arthritis is true, judicious use of antibiotics might be
one of the few ways we have to prevent JIA and other autoimmune
diseases.
Janani Sankar
Department of Pediatrics,
Kanchi Kamakoti CHILDS Trust Hospital,
Chennai, India.
Email: [email protected]
|
|
|
|