Neonatal urinary tract infection may increase
the risk
of childhood asthma
C.-H. Lin1,2 & Y.-C. Wang3,4 & W.-C. Lin4,5 & C.-H. Kao6,7
Introduction
Asthma is the most common chronic respiratory inflammatory disease in childhood, and an increase in the prevalence of asthma in the latter half of the 20th century has been reported [1, 2]. Numerous studies have investigated possible explanations for this increase, but the causes of asthma remain unclear. Increasingly more studies have focused on early-life
events, such as respiratory viral infections, that may increase the risk of childhood asthma [3–5]. Other perinatal risk factors— including babies born preterm, of low birth weight, with
respiratory distress syndrome or neonatal jaundice, and
needing positive pressure ventilation at birth—increase susceptibility to asthma [2, 6, 7]. Numerous studies have suggested
that early exposure to infections in utero or early in
life is a potentially critical risk factor in the development of allergic disease [6, 8–10]. However, little is known about whether neonatal nonrespiratory bacterial infections, such as neonatal urinary tract infection (UTI), are associated with the risk of childhood asthma.
The National Health Insurance Research Database (NHIRD) is a representative database that includes basic patent data and all medical records and accurately reflects an objective profile of morbidity in Taiwan [11]. We hypothesized that newborns
with UTI have an increased risk of childhood asthma compared with those without UTI. To test this hypothesis, we conducted a
population-based retrospective cohort study by analyzing the NHIRD.
Methods and materials
The National Health Insurance (NHI) program, launched in 1995, is the largest health insurance program in Taiwan, and the coverage rate has been nearly 99 % of the resident population of Taiwan since 2000. Following Taiwan’s implementation
of its NHI program, the NHIRD was established by the National Health Research Institutes (NHRI) in 1996. Before the National Health Insurance Research Database (NHIRD) was released for research purposes, the identification of each insurant was encrypted to protect personal privacy.
This study was based on claims data information from a random sample of all insured children under 18 years of age in Taiwan for the 1996–2008 period, which is a subset of the NHIRD. This dataset contains registration files and original claims data for reimbursement, such as information concerning demographic characteristics and diagnosis of disease for each beneficiary.
Study population
We used the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) to diagnose the disease of each beneficiary in a retrospective cohort study. Several Taiwan studies had demonstrated the high accuracy and validity of ICD-9 diagnosis in NHIRD. Taiwan launched a national health insurance (NHI) in 1995, operated
by a single-buyer, the government [7, 12–14]. Medical reimbursement specialists and peer review should scrutinize all
insurance claims. The diagnoses of urinary tract infection and asthma were based on the ICD-9 codes which were judged and determined by related specialists and physicians according to the standard criteria. Therefore, the diagnoses and codes for urinary tract infection and asthma used in this cohort study should be correct and reliable.
For enhancing the diagnostic certainty, UTI was defined as the patient having at least one hospital admission or at least two visits for outpatient medical services related to UTI. We established a UTI cohort as patients with newly diagnosed UTI (ICD-9-CM: 599.0 or 771.82) and aged under 1 month for the period 2000–2006 (n=3,501). The clinical diagnosis
four-foldfrequency-matched the non-UTI participants according to sex, urbanization of residential area, parental occupation, and index dates, to select a comparison cohort (n=13,992). According to the level of population density (people/km
[2]), proportion of higher education, size of elderly and agricultural populations, and number of physicians per 100,000
people, we classified urbanization into seven ordered levels (level 1 to 7). Because of the relatively small number of level 6 and level 7 classifications, we combined them with level 5. Level 1 demonstrated the highest levels of urbanization, whereas level 4 was the least urbanized. All of the participants with a history of asthma before the index date, older than 1 month, with vesicoureteral reflux (VUR; ICD-9-CM: 5,937) diagnosis or missing baseline data on birthday or sex were excluded. Regarding potential risk factors for adjustment, we considered not only demographic characteristic factors (sex,
urbanization of residential area, and parental occupation) but also the comorbidity history in both cohorts. A comorbidity date before the index date was defined according to infections (ICD-9-CM: 771), neonatal jaundice (ICD-9-CM: 774), preterm low birth weight (ICD-9-CM: 764 and 765), and other
fetal and newborn respiratory conditions (ICD-9-CM: 770). Objectives
In this study, our major focus was measuring the occurrence of asthma (ICD-9-CM493).We calculated the follow-up time in person-years for each participant from the index date until the diagnosis of asthma, the end of 2008, or withdrawal from the insurance system (because of death or loss to follow-up). Data availability statement
All data and related metadata were deposited in an appropriate public repository. The data on the study population
that were obtained from the NHIRD (http://w3.nhri.org.tw/ nhird//date_01.html) are maintained in the NHIRD (http:// nhird.nhri.org.tw/). The NHRI is a nonprofit foundation established by the government.
Ethics statement
The NHIRD encrypts patient personal information to protect privacy and provides researchers with anonymous identification
numbers associated with relevant claims information, including sex, date of birth, medical services received, and prescriptions. Patient consent is not required to access the NHIR
D. This study was approved by the Institutional Review Board (IRB) of China Medical University (CMU-REC-101-012).
The IRB specifically waived the consent requirement. Statistical analysis
Baseline characteristics of the demographic distributions of sex, urbanization of residential areas, parental occupation,
and comorbidity history in the UTI and non-UTI cohorts were described to analyze the differences in these variables between
UTI and non-UTI cohorts by employing a chi-squared test. When the assumption of a chi-squared test was relaxed, we used Fisher's exact test to assess the differences for variables, such as comorbidity of VUR, in the two cohorts. The incidence rate was calculated in both cohorts and stratified according to demographic variables and comorbidity history, to measure the incidence rate ratio (IRR) of asthma in these variables by using Poisson regression analysis. The adjusted hazard ratio (aHR) of asthma and 95 % confidence interval (95 % CI) were estimated using the multivariable Cox proportional hazards model to measure the association between UTI and asthma. Kaplan-Meier cumulative incidence plots in the UTI and non-UTI cohorts were generated showing the amount of time before asthma for all end points, and a log-rank test was employed to assess the differences between the two cumulative incidence curves. Analyses were performed using the
SAS 9.3 statistical package (SAS Institute Inc., NC, USA), with P<0.05 in two-tailed tests considered significant.
Results
During the study period, 3312 children with UTI and 13,243 children without UTI were characterized (Table 1). All of the children were younger than 1 month. The UTI and non-UTI cohorts exhibited similar distributions of demographic variables, but a history of comorbidity was more common in the
UTI cohort (P<0.0001). There were no participants with VUR in both cohorts.
measured in the UTI and non-UTI cohorts. The overall asthma incidence rate was 1.53-fold significantly higher in the UTI cohort than in the non-UTI cohort (70.3 vs 45.8 per 1000 person-years). After we adjusted for potential risk factors, the overall risk of asthma remained higher in the UTI cohort (aHR=1.47, 95 % CI=1.35–1.59). The incidence rate was higher in boys than in girls. For boys, the aHR of asthma was 1.48 (95 %=1.34–1.64) higher for UTI patients than for non-UTI patients. Compared with the non-UTI cohort, the UTI patients’ aHR of asthma was 1.44 (95 % CI=1.21– 1.72) for level 1 urbanization, 1.49 (95 % CI=1.28-1.72) for level 2 urbanization, 1.40 (95 % CI=1.17–1.69) for level 3
urbanization, and 1.55 (95 % CI=1.33–1.81) for level 4 urbanization. An occupation-specific analysis revealed an incidence
of asthma in both cohorts, regardless of parental occupation; however, patients with UTI had a greater risk of asthma than those without UTI. Furthermore, the incidence of
asthma for patients with comorbidity was higher than that in patients without comorbidity. The aHR of asthma in the UTI cohort with comorbidity was 1.32 (95 % CI=1.16–1.51)
higher than that of the non-UTI cohort with comorbidity. Figure 1 illustrates that the cumulative incidence of asthma
was 9.3%higher in the UTI cohort than in the non-UTI cohort during a 9-year follow-up period (38.1 % vs 28.8 %; log-rank test <0.0001). Table 3 shows that the IRR of asthma in the UTI cohort and the non-UTI cohort declined over time. Before the 5-year follow-up, the aHR of asthma was significantly higher in the UTI cohort than in the non-UTI cohort (aHR=2.00 for <1-year follow-up, aHR=1.45 for 1–3 years follow-up, aHR= 1.33 for 3–5 years follow-up).
Discussion
Based on our research, this is the first population-based retrospective cohort study to demonstrate an association between
neonatal UTI and the risk of childhood asthma. This association was independent of other risk factors including age, sex, urbanization, occupation, and comorbidity (adjusted for HR). The hygiene hypothesis suggests that decreased microbial exposure in childhood leads to an increasing prevalence of
UTI and childhood asthma; in other words, a multitude of
infectious agents can induce protection from various immunological disorders [15]. Burgess et al. demonstrated that
childhood infectious disease protected against asthma persisting in later life but that pertussis and measles were
associated with new-onset asthma after childhood [16]. In addition, Illi et al. reported that endotoxin in a child’s mattress
was inversely associated with atopic sensitization and asthma [10]. Both of these studies have found that childhood infection agents, but not neonatal infection, had a protective effect against allergic disease.
In a large population-based study, Algert et al. showed that, as in utero exposure to both UTI and preterm, prelabor rupture of membranes carries an increased risk of childhood asthma, immune system response rather than a specific organism [17] may generally be the relevant factor. The findings of
McKeever et al. are also inconsistent with the “hygiene hypothesis” [8]. Their study demonstrated sufficient statistical
power to examine the effects of 14 types of infection, but without UTI in 6-month intervals from birth. However, none of these personal infections appeared to protect against the
development of allergic disease consistently, even after adjustment for consulting behaviour. They also found no consistent
evidence that antibiotics increased the risk of developing allergic disease, although there might be an increased risk of
early diagnosis of allergic disease with the use of antibiotics. Montgomery et al. reported that newborns who spent their first night in a communal nursery were at increased risk of developing hay fever because infants in the nursery were more likely subjected to low doses and short durations of nonfamilial microorganisms [18]. The authors speculated that the development of the immune system is strongly influenced by
early exposure to infectious agents, potentially leading to dysfunction.
Previous studies have reported that impaired innate immune factors may predispose to infection, particularly rhinovirus, Streptococcus pneumoniae, Salmonella enteritidis, and Escherichia coli [19, 20]. The mechanism in effect indicates
that allergic sensitization can induce similar impairment of innate immunity through toll-like receptor signal transduction.
Conversely, the infection destroyed the innate immune system, triggered impaired type 1 helper T cells, and induced a sensitization reaction of asthma. This may explain why neonatal urinary
infection may present a risk factor for childhood asthma, because these microorganisms may destroy the naïve innate immune system and play a role in sensitization to asthma. The merit of this study is that we show an association between UTI and subsequent risk of childhood asthma by using a large population database with minimal selection bias. However, there are some limitations to our study. First, the diagnosis of asthma was based on ICD codes; hence, detailed laboratory data and clinical information, such as C-reactive protein, urine culture, the IgE level, specific IgE, and severity of asthma, was lacking. Second, the data about other potential risks or protective factors beyond the neonatal age are unavailable in the NIRHD. Third, the participants in this study were
ethnically Chinese; thus, the results might not be applicable to other populations. Fourth, follow-ups do not extend beyond a participant’s eighth birthday. Finally, the use of antibiotics in early infancy has been observed to substantially increase the risk of asthma between the ages of 5 and 10 [21]. There is still no evidence showing that the use of antibiotics in newborns causes a risk of childhood asthma; therefore, confounding risks associated with antibiotics or UTI will require careful control study.
In conclusion, neonatal UTI may increase subsequent risk of childhood asthma, and that the influence is stronger in children younger than 5 years old. Future research on the common immunological aberrancies of these disorders is warranted.
