Prevalence, time trends and treatment practices of asthma in India: the Global Asthma Network study

Discussion

The GAN Phase I study was the largest multicentre study since the ISAAC Phase III to analyse the prevalence of asthma symptoms in children and adults in India. The prevalence of current wheeze was 3.16%, 3.63% and 3.30% in the 6–7 years, 13–14 years and adults, respectively. The comparison of prevalence of current wheeze across Indian centres in ISAAC Phase I and Phase III, and GAN Phase I, showed a significantly lower prevalence in both the age groups of children in the GAN study (p<0.001), but a higher prevalence for nocturnal cough (p<0.001). Most subjects with symptoms of asthma were not diagnosed clinically with asthma. A significantly higher number of current wheezers/symptoms of severe asthma with a doctor's diagnosis of asthma were taking either inhaled medications or inhaled steroids or oral medications in all three groups (p<0.001) compared to those without a doctor's diagnosis of asthma.

The time trends of prevalence of asthma from our study are contrary to the widely held belief that asthma is increasing in recent decades. However, our study reveals a significant reduction in the prevalence of current wheeze, which has been considered a symptom of asthma. The methodology, recruitment sites and investigators were the same in the ISAAC and GAN studies, thereby allowing for a reasonable comparison of data to assess the time trends. A significant reduction was also noted in the GAN study in the prevalence of most of the causal factors of current wheeze as compared to the previous ISAAC studies, such as paracetamol use, maternal smoking, farm exposure, pets in house, trucks passing outside the house and antibiotic use during the first year of life (table 2). In India strict vehicle emission norms have also been implemented during this period. Conversely, use of factors such as fruit consumption and child ever breastfed were increased significantly in the GAN Phase I study as compared to the ISAAC Phase III study, which are considered as protective factors for asthma [11]. National asthma guidelines were also formulated and initiated, which would have also resulted in better outcomes [15]. The change in aforementioned environmental factors may be a possible explanation for reduction in the prevalence of current wheeze.

Nevertheless, there is a wide variation in the prevalence of wheeze in comparative studies. A questionnaire-based study in adults in India revealed the prevalence of asthma as 2.05% [3]. Another multicentre research study has reported a prevalence of asthma as 2.38% in adults [16]. An ISAAC questionnaire-based study reported the prevalence of bronchial asthma as 13.1% in children aged between 11 and 16 years (n=927). The prevalence of current subjects with asthma (defined as asthma episode(s) in the past 12 months) was 10.2% [17]. Pal et al. [18] profiled 15 epidemiological studies related to asthma in children and reported the prevalence of asthma as 2.74%. Significant time has elapsed since these studies were conducted and published. These studies also had varied methodology, time frames and study population, which could be possible reasons for this difference. The recently published GBD data estimated 34.3 million cases in an Indian population of 1330 million, almost comparable to our study [4].

The GAN Phase I global data also show significant changes as per centre, age and income in prevalence of asthma across the world [19]. The world total data showed a significant decline in current wheeze, with both increase and decrease noted as per individual centre. Preliminary data from few other global centres have shown mixed results regarding time trends in the prevalence of current wheeze [20–22]. Data from a single centre in Mexico showed a 7.9% increase in the prevalence of current wheeze in both age groups compared with ISAAC Phase III [20]. On the other hand, the GAN Phase I study conducted in Bangkok suggests a similar prevalence of current wheeze in the younger age group and slightly lower prevalence in the older age group when compared with ISAAC Phase III [21]. Interestingly, lower prevalence of current wheeze was reported from the centres from low-income group countries while it did not change for high-income countries in either age group. However, prevalence of current wheeze was increased in upper middle-income countries [19]. There have been changes in the micro-environment and macro-environment between the phases of studies across the centres, but the exact cause of the trend would be difficult to identify.

The prevalence of current wheeze varied from centre to centre and was consistent with the previous ISAAC studies. Similar heterogeneity was documented in previous ISAAC studies, with intercontinental, intercountry, inter-regional and intra-regional variations [2, 5, 6]. The various cities across India have varying geography, climate, environment and socioeconomic conditions. The lowest prevalence of current wheeze was documented in Bikaner (0.35%) and the highest in Chandigarh (8.54%) in the 6–7-year age group. Bikaner is a desert area located in the western part of India and it experiences dry and hot weather associated with dust storms. Chandigarh, on the other hand, has a cooler and humid climate, situated in the northern part of the country. In the 13–14-year age group, the lowest prevalence was in New Delhi (0.89%) and the highest in Jaipur (6.62%) (supplementary tables 1 and 2, supplementary figure 1). Among adults, New Delhi (0.88%) reported the lowest prevalence and Kottayam (6.02%) the highest prevalence (supplementary table 3). The findings are contradictory as Jaipur is located in western India and has a dry and arid desert climate while New Delhi is situated in the northern part of the country, with poor air quality. Kottayam is situated in southern India with a warm and tropical climate. Such variation is also observed in global data, which is difficult to explain. After ISAAC Phase III, compressed natural gas (CNG) replaced petrol and diesel in nearly all public vehicles in New Delhi in December 2002, including truck traffic [23]. Diesel exhaust particles have been associated with an increase in Th2 and Th17 cytokines, which are responsible for the asthmatic response [24]. Reduction in diesel exhaust particles may in turn have been a contributing factor for a decrease in prevalence in New Delhi. The INSEARCH study also reported variation in the prevalence of bronchial asthma (adults) among the 12 participating centres (from 0.4% in Secunderabad to 4.8% in Kolkata) [3].

The self-reported prevalence of wheeze was significantly more common in boys in both age groups (p<0.001) [25]. It is unclear why there is this gender difference in children; possible explanations are variations in sex hormone levels [26], bronchial lability [27] and allergen sensitivities [28]. Interestingly, among adults this difference was not noted for current wheeze and a significantly higher prevalence of asthma ever and doctor-diagnosed asthma were seen in women, which is consistent with previous reports in adults. Previous health surveys conducted from 2005 to 2006 in adults in India have reported a prevalence of wheeze of 1.8% (95% CI 1.6–2.0) and 1.9% (95% CI 1.8–2.0) in men and women, respectively [29].

Nocturnal cough is a symptom of asthma that has increased in prevalence during GAN Phase I compared to ISAAC Phase I and Phase III. However, nocturnal cough can be attributed to other causes also such as gastro-oesophageal reflux disease, post-nasal drip, bronchiectasis and medication-induced [30]. It is a non-specific symptom and all nocturnal cough would not be due to bronchial asthma. Nonetheless, it is worth noting the significant increase in the prevalence of nocturnal cough; a similar increase was also noted in both the age groups in the worldwide GAN Phase I data, a trend that is difficult to explain [19].

Our study reveals a gap in the diagnosis of asthma. Almost 82% of current wheezers and 70% of subjects with symptoms of severe asthma were not clinically diagnosed as having asthma by a doctor. This gap in diagnosis is high compared to a similar study that reported almost a third of the subjects remained undiagnosed [31]. Under-treatment is another problem; among doctor-diagnosed subjects with wheeze and with symptoms of severe asthma, daily ICS were used in less than 9% among different age groups of subjects (table 5). Under-diagnosis and under-treatment may be attributed to lack of medical facilities, poverty, illiteracy, ignorance on the part of patients, improper techniques of medication intake, non-adherence and poor communication skills of medical personnel [32–34]. Apart from these factors an interesting aspect worth discussing is disease terminology used by asthmatic patients and their treating doctors. When an asthmatic patient consults a doctor, only 71% of doctors refer to “asthma” as the name of their disease and 29% use other terminology. The problem is worse at patient level; only 23% of asthmatic patients call their disease “asthma” while the rest use terminology like swas, dama or cold and cough [35]. Asthma is considered to be a stigma and many parents conceal the disease and therefore avoid medications or give only when a patient is symptomatic or unable to tolerate symptoms. Incorrect notions such as that inhalers are harmful and habit-forming also play a role in treatment non-adherence. Regular use of ICS among undiagnosed current wheezers was thereby very low; under-treatment may account for high school absenteeism, urgent medical consultation to a doctor and emergency department visits (table 3). Even hospitalisation due to breathing problems was significantly high. Similarly, indexes of uncontrolled asthma such as exercise-related wheeze in children and restriction of usual activities were also significantly higher in current wheezers than healthy subjects (table 3). These findings suggest high morbidity and use of hospital resources among current wheezers.

Untreated children with asthma have been found to have higher school absenteeism in similar studies [36]. Exercise-induced asthma, emergency visits, hospitalisation and school absenteeism are markers of uncontrolled asthma [37]. However, all these markers were more significantly affected in subjects having a clinical diagnosis of asthma, probably indicating that more severely affected persons consult a doctor more frequently and thereby have a higher chance of getting clinically diagnosed with asthma. However, earlier studies show that more than half of asthmatic subjects discontinue asthma medications once they tolerate symptoms and this tendency of symptom tolerance persists in more than a third of patients, even after education [32, 33]. This issue needs to be addressed in asthma education programmes as it leads to significant clinical implications in the form of non-compliance.

Our study thus highlights an unmet need in terms of diagnosis and treatment of symptoms of asthma, thereby leading to significant morbidity resulting from the disease. Lack of patient education, concerns about medication side-effects and the stigma associated with inhalation devices might also contribute to this skewed treatment approach [32, 34].

Underdiagnoses, delayed treatment, under-treatment or no treatment may be responsible for increased morbidity due to asthma in terms of DALYs and disproportionately high mortality [1, 4]. These findings emphasise the need for national-level health programmes to address better strategies to diagnose and manage asthma, and also the need to communicate a diagnosis of asthma to the patient.

The GAN Phase I study was a self-reported, questionnaire-based study; objective measures of assessment of asthma and its severity, such as spirometry, were not used for diagnosis of the disease. Conducting spirometry for every child across all centres would not have been logistically feasible and would not have allowed for comparison with earlier studies. Because it was a questionnaire-based study, answers were subject to the understanding of the children and adults. Adolescents may have a poor perception of their symptoms; thus questionnaire-based assessment tools may underestimate the actual prevalence of the illness. Cough is a common symptom in subjects with asthma, especially children. Although wheeze is more specific for a diagnosis of asthma, its sensitivity is low and may therefore underestimate the prevalence of asthma. However, the previous ISAAC studies have used this particular question to calculate the prevalence of asthma [2, 5, 6]. The validated ISAAC questionnaire is the most authentic means of collecting childhood asthma prevalence globally [14, 38]. Furthermore, because the study aimed to compare the current prevalence of wheeze with the previous ISAAC data, a similar questionnaire was used to maintain consistency. The study was limited by the fact that the centres participated voluntarily, and were not randomly selected. Although the participating centres were voluntary they represented the various geographies including north, south, east, west and central. Further, each centre was divided into four zones, and a fraction of schools were randomly selected from each zone to ensure appropriate representation. This same methodology was adopted by the GAN Phase I study worldwide and the previous ISAAC studies as well. The study population was representative of the population of school-going children of the respective nine centres. Despite these steps to strive for equal representation from across centres, the prevalence of asthma varies across rural and urban areas, and within cities itself due to differences in socioeconomic conditions, thus to generalise the results to the entire country or the world would not be appropriate. Potential bias could exist as some children may have refused to participate and all children may not be attending school. Furthermore, not all schools across the country were covered. Recall bias could also exist, as exposures in the first year of life could have been forgotten or incorrectly remembered. However, the same protocol was used in the previous ISAAC studies, thereby allowing for valid comparison of time trends. Prevalence of current wheeze in adults could be obtained for only those adults (age group 31–44 years) whose children were in a particular age group and were attending the participating schools; thus, a selection bias may exist.