Multivariate modelling of responses to conditional items: New possibilities for latent class analysis

Questionnaire data may contain missing values because certain questions do not apply to all respondents. For instance, questions addressing particular attributes of a symptom, such as frequency, triggers or seasonality, are only applicable to those who have experienced the symptom, while for those who have not, responses to these items will be missing. This missing information does not fall into the category ‘missing by design’, rather the features of interest do not exist and cannot be measured regardless of survey design. Analysis of responses to such conditional items is therefore typically restricted to the subpopulation in which they apply. This article is concerned with joint multivariate modelling of responses to both unconditional and conditional items without restricting the analysis to this subpopulation. Such an approach is of interest when the distributions of both types of responses are thought to be determined by common parameters affecting the whole population. By integrating the conditional item structure into the model, inference can be based both on unconditional data from the entire population and on conditional data from subjects for whom they exist. This approach opens new possibilities for multivariate analysis of such data. We apply this approach to latent class modelling and provide an example using data on respiratory symptoms (wheeze and cough) in children. Conditional data structures such as that considered here are common in medical research settings and, although our focus is on latent class models, the approach can be applied to other multivariate models. Copyright © 2009 John Wiley & Sons, Ltd.     

Adipokines in cord blood and risk of wheezing disorders within the first two years of life

BACKGROUND: Adipokines are involved in the regulation of many inflammatory processes and are present at very high concentrations in cord blood of term infants. OBJECTIVE: We analysed data of a large prospective birth cohort study to examine whether adiponectin and leptin concentration in cord blood are determinants of wheezing disorders in children within the first 2 years of life. METHODS: Seven hundred and forty mothers and their newborns were included in this analysis. Adiponectin and leptin concentrations were measured in cord blood. The cumulative incidence of physician-reported asthma or obstructive bronchitis was recorded during a 2-year follow-up. RESULTS: During the first 2 years of life, asthma or obstructive bronchitis was reported by the caring paediatricians for 157 (19.6%) of the children. We found a strong interaction of cord blood adiponectin and history of atopic disease in the mother with respect to the risk of physician-reported asthma or obstructive bronchitis (P=0.006). Compared with children with cord blood levels in the middle quintile (reference category), the odds ratios for physician-reported asthma or obstructive bronchitis in the bottom quintile and top quintile were 0.14 [95% confidence interval (CI) 0.02-0.90] and 2.12 (95% CI 0.67-6.66), respectively (P for trend=0.0003), among children of mothers with a history of atopy. This association was independent of other established risk factors. Leptin levels in cord blood were not associated with risk of asthma or obstructive bronchitis. CONCLUSIONS: In children of mothers with a history of atopy, concentrations of adiponectin in cord blood could play an important role in determining risk of wheezing disorders in early childhood.     

Are chronic wheezing and asthma-like attacks related to FEV1 decline? The Cracow study

Analysis is presented of the relationship between decline in ventilatory function and possible symptoms of bronchial hypereactivity. Data comes from three consecutive surveys within a 13-year follow-up study of a population sample of 1747 persons (718 males and 1029 females) who completed spirometric testing and interviews. Chest symptoms recorded in the two initial surveys were used as baselines for the prognostic study on FEV₁ change. To evaluate changes in lung function the FEV₁ decline index, expressed as percentage of FEV₁ loss during the study period, was calculated. Chronic wheezing and asthma-like attacks were very strongly related to FEV₁ baseline level. Multiple linear regression analysis for each separate symptom including age and smoking habits as confounders and showed that the FEV₁ decline index was higher and more significant when wheezing and asthma-like attacks or chronic cough were present. When the combination of symptoms were introduced into the regression model, the effect of wheezing became insignificant. Only attacks of shortness of breath in both sexes and chronic cough in women were significantly associated with more rapid FEV₁ decline. However, the effect of persistent wheezing in men had a rather substantial impact on FEV₁ change. Comparing the effect of the number of cigarettes smoked with the effect of symptoms, one can see that FEV₁ decline due to attacks of shortness of breath is comparable to that seen with heavy smoking. There was no strong evidence that the symptoms under study precede the more rapid FEV₁ decline.     

Does previous infection protect against atopic eczema and recurrent wheeze ininfancy?

BACKGROUND: Frequent infection in infancy and early childhood has been hypothesized to explain the low prevalence of asthma and other atopic disease among children in developing countries (the so-called 'hygiene hypothesis'), but the low prevalence in Eastern Europe remains unexplained. OBJECTIVE: To test the hygiene hypothesis in the Republic of Belarus by examining the relationship between gastrointestinal (GI) and respiratory infection and two potentially atopic outcomes in the first 12 months of life: atopic eczema and recurrent wheeze. METHODS; We carried out two case-control studies nested within a large (n=17 046) randomized trial in Belarus, with cases defined as (1) first occurrence of atopic eczema (n=819) and (2) second episode of wheezing (n=112). Incidence density sampling was used to select four matched controls born within 1 month at the same hospital as the case. Exposure was defined as one or more episodes of GI or respiratory infection (examined separately) with onset >7 days before onset of the case's atopic outcome. Analyses controlled for family atopic history, duration of exclusive breastfeeding, sex, birth weight, maternal education, and (for recurrent wheeze) maternal smoking. RESULTS: For atopic eczema, prior GI infection occurred in 7.4% of cases vs. 6.0% of controls [adjusted OR=1.27 (0.94-1.72)] and prior respiratory infection in 35.2% vs. 32.6% [adjusted OR=1.14 (95% CI=0.94-1.37)]. For recurrent wheeze, prior GI infection occurred in 9.8% of cases vs. 7.4% of controls [adjusted OR=1.30 (0.60-2.82)]. CONCLUSION: Our results do not support the hypothesis that infection protects against atopic eczema or recurrent wheezing in the first 12 months of life.     

Latex allergy in infants younger than 1 year

BACKGROUND: The prevalence of latex allergy in children is increasing worldwide. Previous multiple operations or atopic predisposition are known risk factors. In contrast, only sporadic cases of latex allergy have been reported in infants younger than 1 year, and the causative latex-containing products or symptoms in young infants have not been studied in detail. OBJECTIVE: The purpose of this study is to analyse the symptoms and risk factors of latex allergy in young infants. METHODS: Cases of latex allergy in infants younger than 1 year were studied in detail. Clinical course, causative latex-containing products were spotted and detailed analysis for latex allergy in patients and patients' parents was performed. CONCLUSION: We report nine cases of latex allergy in infants younger than 1 year. None of them have any abnormality or previous operations. Six patients had atopic eczema/dermatitis syndrome, one patient had bronchial asthma, whereas two patients had no overt allergic diseases. Symptoms of latex allergy were wheezing, swelling of face or lips, facial rash, or anaphylaxis, and causative latex-containing products were teat, pacifier, nose cleaner, teether, balloon, or enema tube. All of the nine patients had positive skin prick test to latex and extract from causative latex-containing products, whereas eight patients had positive serum latex-specific IgE. Study for family history revealed that latex allergy was noted in either father or mother in six patients, in both father and mother in one patient, whereas no latex allergy was noted in parents in two patients. It should be noted that all of these patients had latex-induced symptoms at home. Latex allergy in young infants may not be unusual. Physicians should be aware of latex allergy, and care should be taken to avoid contact with latex in young infants, especially when there is family history for latex allergy.     

Risk factors for wheezing in Ukrainian children: Ukraine European Longitudinal Study of Pregnancy and Childhood Group

The prevalence of wheezing in children varies widely around the world. The reasons for this geographic variability remain unclear but may be related in part to exposures in the home environment during pregnancy and early childhood. We investigated the prenatal and early childhood risk factors for wheezing symptoms among 2127 children aged 6–8 years who were participants in the Ukrainian component of the European Longitudinal Study of Pregnancy and Childhood (ELSPAC). Cases included the 169 children whose parents answered yes to the International Study of Asthma and Allergy in Children (ISAAC) question: 'Has your child had wheezing or whistling in the chest in the past 12 months' during the ELSPAC assessment of the children at age 7. These were compared with the 1861 children in the cohort whose parents answered 'no' to this question.
Factors significantly associated with increased risk of wheezing illness at age 7 in adjusted analyses included mother's asthma [adjusted odds ratio (OR) 3.46, 95% confidence interval (CI) 1.22, 9.85]; mother's allergy problems (OR 1.43, [1.00, 2.05]); rarely playing with other children at age 3 (OR 1.84, [1.09, 3.11]); water intrusion (OR 1.62, [1.09, 2.39]) and inadequate heating of the home (OR 1.52, [1.06, 2.16]) during pregnancy. Factors protective of wheezing at age 7 included being first-born (adjusted OR 0.70, 95% CI 0.50, 0.98); living in the city of Dniprodzerzynsk as compared with Kyiv (OR 0.36, [0.24, 0.54]) and weekly contact with furry animals (OR 0.44, [0.20, 0.97]) before age 3. The constellation of risk factors for wheezing in Ukrainian children is similar to that of children in other parts of the world. Known risk factors do not account for the significant between-city variability of wheezing in Ukrainian children.     

Recurrent wheezing in relation to environmental risk factors in infancy

Clinical course and environmental factors were recorded in a prospective study of 276 unselected infants followed from birth to the age of 18 months. The study was performed with a questionnaire at the age of 6 and 12 months and a physical examination at 18 months. Fifty-nine (21%) of the children had greater than or equal to 2 episodes of wheezing before they were 18 months old. A total of 58 (21%) of the children belonged to the lowest social class V, 182 (66%) were daily exposed to passive tobacco smoking at home and/or in daycare, 164 (59%) were breastfed greater than or equal to 3 months, 192 (70%) were in daycare, 62 (22%) lived in flats and 167 (61%) were in daily contact with furred pets at home and/or in daycare. In social class V a preponderance of children were exposed to passive tobacco smoking, a majority were living in flats and a minority were breastfed greater than or equal to 3 months. Linear logistic regression analysis was used for the purpose of assessing the causal effect of environmental risk factors on the risk of recurrent episodes of wheezing before the age of 18 months. The study demonstrated that male sex and daily exposure to passive tobacco smoking were significant risk factors with estimated odds ratios 1.9 and 2.4, respectively. Maternal tobacco smoking seemed to be associated with the highest risk. There was a tendency--though not significant--indicating that breastfeeding greater than or equal to 3 months had a protective effect.(ABSTRACT TRUNCATED AT 250 WORDS)     

Wheezing at 8 and 13 years: changing importance of bronchiolitis and passive smoking

A group of 153 children (51 with a history of bronchiolitis and 102 matched controls) were evaluated in a historical cohort study at a mean age of 8 years and again at 13 years to test the primary hypothesis that mild bronchiolitis, far more common than severe (hospitalized) bronchiolitis, predicts wheezing. A secondary hypothesis was that passive smoking also predicts wheezing. Many potentially confounding variables such as family history of asthma were controlled in analyses. Analysis at 13 years produced results that were not anticipated from previous analysis of interviews at age 8. Although mild bronchiolitis was a powerful predictor of wheezing at age 8 years, it was no longer a strong predictor of wheezing at age 13 in either bivariate or multivariate analysis. Although epidemiologic studies, by their nature, cannot prove causality, findings are consistent with the hypothesis that sequelae often follow mild bronchiolitis but diminish during childhood. Maternal smoking was a powerful predictor of wheezing at age 13 in bivariate analysis (Kendall's Tau B = 0.19, P less than 0.01) and in multivariate analysis (odds ratio = 2.67, P less than 0.01). In children at highest risk for wheezing, males with a family history of asthma, multivariate analysis suggested that maternal smoking is associated with an increase in wheezing from 36% to 60%. We conclude that passive smoking, previously identified as a risk factor in this population for both bronchiolitis in infancy and wheezing at age 8, is a risk factor for wheezing-associated morbidity throughout the childhood years.     

Prenatal paracetamol exposure and risk of asthma and elevated immunoglobulin E in childhood

BACKGROUND: We recently found that paracetamol (acetaminophen) use in late pregnancy was associated with an increased risk of early wheezing in the offspring. OBJECTIVE: To see whether use of paracetamol in late pregnancy is associated with an increased risk of asthma, wheezing and other atopic outcomes in the child at school age. METHODS: In the population-based Avon Longitudinal Study of Parents and Children, we measured associations of paracetamol and aspirin use in late pregnancy (20-32 weeks) with asthma, hayfever, eczema (n = 8511) and wheezing (8381) in the offspring at 69-81 months, and with atopy (positive skin prick test to Dermatophagoides pteronyssinus, cat or grass, n = 6527) and blood total IgE (n = 5148) at 7 years. We used logistic and linear regression to analyse binary outcomes and log-transformed IgE, respectively, controlling for potential confounders. RESULTS: Use of paracetamol, but not aspirin, in late pregnancy was positively associated with asthma (odds ratios (ORs), comparing children whose mothers took paracetamol 'sometimes' and 'most days/daily' with those whose mothers never took it, 1.22 (95% confidence interval (CI): 1.06-1.41) and 1.62 (95% CI: 0.86-3.04), respectively; P trend = 0.0037), wheezing (ORs 1.20 (95% CI: 1.02-1.40) and 1.86 (95% CI: 0.98-3.55), respectively; P trend = 0.011), and total IgE (geometric mean ratios 1.14 (95% CI: 1.03-1.26) and 1.52 (95% CI: 0.98-2.38), respectively; P trend = 0.0034), but not hayfever, eczema or skin test positivity. The proportion of asthma attributable to paracetamol use in late pregnancy, assuming a causal relation, was 7%. CONCLUSION: Paracetamol exposure in late gestation may cause asthma, wheezing and elevated IgE in children of school age.     

Immunoglobulin G subclasses in wheezing infants

Background: The wheezing infant is a common but difficult patient to approach diagnostically. The prevalence of immunoglobulin (Ig) G subclass deficiency in wheezing infants is still controversial. Methods: We studied the serum concentration of IgG subclasses in 38 wheezing infants (aged6–24 months) who had not received systemic steroids before investigation and in 30 healthy age matched controls6–31 months). Results: The prevalence of one or more IgG subclass deficiencies was 3 1.6% in wheezing infants and 26.7% in controls. There was no significant difference in prevalence of IgG subclass deficiency between patients and controls (P > 0.05). The mean concentration of IgG subclasses in patients were compared with controls. There was no significant difference in mean serum concentration of IgG₁, G₂ and G₃ subclasses. However, there a trend towards higher concentrations of IgG₄ in wheezing infants and this difference for IgG₅ was significant (P < 0.01). Immunoglobulin G subclass deficiency was found in 25 and 36.4% of wheezing infants who experienced from two to four and five or more wheezing episodes in 2 years, respectively (P > 0.05). Conclusion: Our findings suggest that wheezing in infancy is not associated with IgG subclass deficiency, and in wheezing infants low IgG subclasses levels do not increase the frequency of wheezing. However, is a relationship between recurrent wheezing and serum IgG₄ subclass concentration.