In this retrospective cohort study we found that allergic rhinitis is strongly predictive of the development of asthma even after adjustment for other risk factors for asthma. Other significant risk factors for the development of asthma were female sex, pet ownership, and family history of allergic diseases. With the exception of Parietaria judaica, sensitization to the allergens tested did not predict long-term asthma diagnosis. In addition, we have shown for the first time that treatment with allergen immunotherapy decreases the incidence of asthma in adults with allergic rhinitis.
The effect of rhinitis on the onset of asthma has been already investigated in longitudinal studies. Huovinen et al.  found that hay fever increased the risk of development of asthma during a 15-year follow-up period by 4 times among adult men and by 6 times among women. However, no information on atopic status was available. Similarly, in the cohort of Brown University freshmen  both allergic rhinitis and positive skin test responses increased the risk of development of asthma by about 3 times. This is in support of our findings, although the risk from our data was much higher. Data from large population-based studies clearly show that rhinitis is a risk factor for asthma among subjects with negative, as well as positive, skin test responses thus suggesting that rhinitis and asthma are not associated simply because they share atopy as a common risk factor [15, 16, 18, 30].
In this study, some of the factors that have been normally considered important risk factors for the development of asthma were examined. As one would predict, the female sex, the presence of pets in the home and a positive family history of allergic diseases were all predictive of new onset asthma in our study population of Sicilians. Our findings are in agreement with the results from most studies of adults showing that asthma is more prevalent in women than in men [31, 32] and that family history of allergy is consistently identified as an important risk factor for asthma . In keeping with the role of indoor allergens from domestic animals as an important risk factor for asthma and asthma-related symptoms [34, 35], we have shown that the presence of pets in the home was significantly predictive of new onset asthma in individuals with allergic rhinitis.
Sensitization to allergen has been shown to be one of the strongest determinants of asthma, and individuals with a predisposition for atopy are at higher risk . Of all known common allergens, the house dust mite is known to be strongly implicated as a potential cause of asthma [37, 38]. However, in this population of Sicilians, house dust mite sensitization was not significantly predictive of new onset asthma. Instead, we have been able to show for the first time that (of all allergens tested) positive sensitization to Parietaria judaica markedly increased the risk of developing asthma. The reasons for this discrepant result are not known, but are probably related to the peculiar characteristics of the inhalant allergen type. The Parietaria pollen is widespread in the Mediterranean area with a very high frequency of sensitization (up to 80% in Sicily) and its long persistence in the atmosphere (Parietaria pollen season in Sicily ranges from February to October) is often responsible for most perennial symptoms . In contrast to mite allergens, Parietaria pollen has very strong allergenic properties and often reaches very high peak levels during its season [39, 40].
It is unclear why a large proportion of individuals with atopy and rhinitis eventually progress to bronchial asthma. Although atopy per se carries an increased risk for subsequent development of asthma in rhinitic individuals, it is likely that chronic exposure to airborne allergens is important. In our study population, up to 70% of rhinitic subjects were allergic to Parietaria pollen. In Sicily, Parietaria-sensitive subjects with allergic rhinitis are likely to be exposed to very high allergen levels, and this high allergenic load may promote progression to bronchial inflammation and asthma. The findings of our recent work in non-asthmatic subjects with allergic rhinitis monosensitized to Parietaria judaica shows a substantial increase in non-invasive surrogate markers of bronchial inflammation during periods of seasonal exposure to Parietaria pollen , thus suggesting that ongoing exposure to Parietaria pollen is closely associated with inflammatory changes in the bronchial airways of subjects with allergic rhinitis, that may advance to clinical asthma.
Another significant finding of the present study is that treatment with allergen immunotherapy reduces the development of asthma in adults with allergic rhinitis. This association has been investigated here for the first time among adults. In support of our findings, clinical research studies have suggested that when allergen immunotherapy is introduced to individuals with allergic rhino-conjunctivitis, the development of asthma may be halted. The pioneering study of Johnstone and Dutton  showed that 28% of children receiving allergen vaccination developed asthma in compared to 78% of placebo-treated children. The Preventive Allergy Treatment (PAT) study in children with grass or birch pollen rhino-conjunctivitis  has been instrumental in providing encouraging evidence to support the notion that specific allergen immunotherapy may stop the development of asthma. From six paediatric allergy centres in Austria, Denmark, Finland, Germany and Sweden, 205 children with moderate to severe hay fever symptoms were randomly assigned either to receive immunotherapy for 3 years, or to an open control group. By the end of the study, the actively treated children developed significantly fewer asthma symptoms. However, in children, wheezing and coughing from non-asthmatic respiratory illness can mimic asthma. The results from the 6 centres were not consistent and there were a small number of children in each centre. Furthermore, 20% of the study population were diagnosed with asthma at baseline and apparently were not excluded. In a recent randomized, placebo-controlled 3-year study of allergen immunotherapy in non-asthmatic, rhinitic adults monosensitized to Parietaria pollen, we reported that 47% of patients in the placebo group developed asthma symptoms by the end of the study, as opposed to only 14% of those treated with immunotherapy . However, the small sample size (n = 29), including only those sensitized to Parietaria, limited our power to detect a statistically significant change between the two groups.
The observed effect of allergen IT in reducing the onset of new asthma cases is of clinical importance. Our study shows that there was a significant 12% reduction in the prevalence of physician-diagnosed asthma in adults with allergic rhinitis, with a number needed to treat (NNT) of 8.7; which is better than that of 10 obtained in the recent Childhood Asthma Prevention Study with omega-3 fatty acid supplementation and house dust mite allergen avoidance . To our knowledge there is no other asthma prevention study that can exhibit a greater effect. Moreover, Parietaria immunotherapy appears to reduce development of asthma even further. Considering the high prevalence of Parietaria sensitivities in our area and the importance of positive sensitization to Parietaria as a major independent risk factor for the development of asthma in this study population, we did secondary analysis for the effects of allergen specific immunotherapy among those individuals with positive sensitivity to Parietaria and those with positive sensitivity to HDM. This analysis clearly shows that treatment with Parietaria immunotherapy reduces development of asthma in adults with allergic rhinitis, with a calculated NNT of 6.6. In contrast, treating positive HDM subjects with allergic rhinitis with HDM immunotherapy failed to reduce the rising incidence of asthma in this subgroup. The observed disparity in responses to different allergen extracts for IT in the present study may provide an additional explanation for the inconsistency in the results from the 6 centres in the PAT-study .
Our study has the advantage of a relatively long follow up period of 10 years. The cohort approach minimizes the possibility of reverse causality that may be encountered in case-control studies, where it is not possible to know if the asthma began after or before the exposure, or in this case allergic rhinitis. Another advantage of this study is the rigorous clinical assessment of asthma diagnosis prior to exclusion at baseline and for its diagnosis as an outcome at the end of follow up. Failing to diagnose actual asthma cases at baseline would have introduced systematic bias, which could affect the results by either increasing or decreasing the observed OR. On the other hand, missing the diagnosis of asthma at the end of follow up would have attenuated the observed OR. Even with all the instruments and expertise available for this study, we had to exclude 39 cases because asthma diagnosis was unclear. The fact that the study subjects were examined by the same respiratory unit at baseline and at follow up 10 years afterwards is important for standardising asthma diagnosis criteria in such a population.
It is also important to obtain an accurate history of steroid and other asthma treatments during follow up. As intranasal use of corticosteroids [44, 45] has been shown to reduce asthma symptoms in patients with allergic rhinitis, we attempted to address these variables in our cohort and there was no difference between asthmatics and non-asthmatics for these treatments. This may explain the higher OR in our study compared to earlier studies.
A possible weakness of our study includes relying on medical records for the selection of the study subjects at baseline. However, all these subjects were examined and carefully diagnosed and documented in the clinic by our specialists. The lower response rate due to the change in the telephone numbers during follow up might have affected our results. However, we do not expect that this lower response is related to the diagnosis of asthma and therefore any random error introduced by lower participation is likely to attenuate rather than exaggerate the observed OR.
In conclusion, our study shows that immunotherapy can be used to reduce progression to asthma later on, especially when there are symptoms of allergic rhinitis and atopy. Allergic rhinitis seems to pose a much higher long-term risk than previously thought. The main allergy-related risk for asthma in Siciliy, and possibly other Mediterranean countries, seems to be Parietaria pollen rather than house dust mites. These are important findings for clinicians to help guide them in the prevention and treatment of their patients. Well conducted clinical trials may shed more light on the significance of our findings in relation to the long-term prevention of asthma.