We found high correlations between impaired asthma-specific quality of life and standard measures of psychological distress, and moderate correlations between impaired asthma-specific quality of life and psychiatric morbidity (i.e., mood and anxiety disorders) in individuals with OA. A cut-off value of < 5.1 on the AQLQ(S)'s emotion subscale could reliably identify individuals with clinically significant levels of depressive and/or anxiety symptoms who need further evaluation by an validated psychiatric interview.
There is limited evidence about the association of psychological stress and asthma morbidity in individuals with OA . When considering the available evidence about the impact of this stress on individuals with non-occupational asthma, we can imagine that an additional psychological burden is associated with OA. This is in accordance with past findings where subjects with OA had slightly but significantly higher impairment in asthma-specific quality of life than those with non-occupational asthma, even when controlling for asthma severity . In a study of asthmatics affiliated with a health maintenance organization in the USA, patients with work exacerbated of asthma had lower quality of life measured according to the mood disturbance, social disruptions, and health concerns subscales of the Mark's Asthma Quality of Life questionnaire, compared to those individuals with no work exacerbated asthma .
Various and probably many unknown factors contribute to impairment of quality of life in individuals with asthma. Malo and co-workers have reported a weak but significant correlation between the original AQLQ with FEV1, bronchial responsiveness, and asthma severity in a more extensive sample of individuals with occupational and non-occupational asthma . The AQLQ(S) used in our study and the original AQLQ questionnaires distinguish themselves on one point: in the AQLQ(S)'s five generic activities (strenuous exercise, moderate exercise, work-related activities, social activities, and sleep) replaced specific activities that could be chosen by the patient in the original questionnaire . We could reproduce these findings and could find a larger correlation of the AQLQ(S) subscales and total score with the objective asthma severity score. We also found a large correlation between the symptom domain of a widely used quality of life questionnaire in chronic obstructive lung disease - the St. George Respiratory Questionnaire, which provides support for quality of life being related to factors other than objective markers of disease severity.
It is currently unknown how treatment of psychological distress or psychiatric morbidity (either using psychotherapy or pharmacotherapy) might affect asthma and psychosocial outcomes in individuals with OA. Disease management programs for major depressive disorders have been shown to be beneficial in reducing the severity of the depression, maintaining employment, increasing short term adherence to medication and improving the individuals quality of life while being cost-effective . It a recent systematic review, Lerner and Henke have shown that individuals with depression have higher unemployment rates, more absenteeism and lower at-work performance than individuals without depression . When on medical leave, individuals with poor mental health are at risk for prolonged work absence . Co-morbid psychiatric disorders are one of the reasons for the adverse socioeconomic outcomes in regards of unemployment and income loss. As these disorders can influence the individual's adherence to medication, lifestyle behaviors such as smoking and managing environmental asthma triggers, it could at least partially explain the persistent symptomatology and bronchial hyperresponsiveness in many individuals with OA seen even years after termination of the exposure to a sensitizing agent [35, 36].
We found medium to large correlations between the individual AQLQ(S) and the PSI. In point-biserial correlations between AQLQ(S) and PRIME-MD outcomes such as having psychiatric disorder the correlations were in the range small to medium. In a population sample in Australia, major depression according to the PRIME-MD was associated with dyspnoea, wakening at night and morning symptoms in asthmatics and these symptoms were shown to have the greatest impact on decrease in quality of life scores in the SF-36 . When using the Hospital Anxiety and Depression (HAD) scale, Rimington and co-workers found moderate correlation of the HAD depression subscale and a somewhat lower correlation of the HAD anxiety subscale with the AQLQ symptoms subscore in a sample of asthmatic patients attending GP offices in the UK . In that study, hardly any correlation of HAD anxiety and depression subscales on the one hand and lung function expressed as forced expiratory volume in one second (FEV1) or peak flow on the other hand could be demonstrated . In contrast, Hommel and co-workers reported anxiety and depression to influence asthma specific quality of life measured with the Living with Asthma Questionnaire (LWAQ). When performing regression analysis, they demonstrated that anxiety had an independent main effect on LWAQ when the model was controlled for depression . The impact of concomitant depression and anxiety seems to be even more deleterious for health related quality in life in individuals with chronic obstructive pulmonary disease .
The AQLQ(S) has been used in a large variety of clinical therapeutic trials and many cross sectional studies on patients with OA. Measuring quality of life with the AQLQ(S) allows to determine the impact of asthma on respiratory symptoms, emotional function, activity limitation as well as environmental stimuli. These factors are important to acknowledge in clinical practice when assessing a patient with asthma. In our sample of individuals with OA who have been removed from exposure to the sensitizing agent, using a cut-off point of 5.1 in the emotional function subscale most reliably distinguishes individuals with significant psychological distress, whereas a cut-off of 4.7 can be used to identify individuals who are at risk of relevant psychiatric disorder according to the PRIME-MD evaluation. It is not the intention of the authors to suggest that the evaluation of patients with the AQLQ(S) emotional subscale can replace a structured diagnostic interview by a psychiatrist - which is considered to be the gold standard - for the diagnosis of psychiatric diseases. Questionnaires such as the Hospital Anxiety and Depression Scale (HADS) have been shown to have a sensitivity of 66-78% and specificity of 83-97% for the diagnosis of either depression or anxiety disorders in a general practice setting . Therefore even administration of tools specifically designed to screen for psychiatric disorders would not allow making an accurate diagnosis and starting treatment without performing a structured psychiatric interview. The advantage of using the AQLQ(S) questionnaire is that it is widely available and regularly used in clinical practice and trials. It can therefore be used as a screening test. Considering the results of the emotional subscale will not only allow to measure the impact of asthma on quality of life but also to identify some individuals in whom a more extensive investigation such as a structured psychiatric interview is warranted. The diagnostic performance of the test using a cut-off of < 5.1 in the emotional subscale score of the AQLQ(S) is modest for identification of clinically important psychological distress according to the PSI. But the performance is less for the diagnosis of current psychiatric disorders according to the PRIME-MD which relies on the diagnostic criteria for depressive and anxiety disorders according to the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, DSM-IV. In fact the positive predictive value for the diagnosis of current psychiatric disease by using a cut-off of 4.7 is close to the predictive value of flipping a coin and would even be lower when this test would be performed in a population with a lower prevalence of psychiatric disorders. There is very limited data suggesting that anxiety and depression are more common in workers in whom the asthma is related to the workplace . In our study all individuals were compensated for OA and one could expect that the prevalence of mood and anxiety disorders is more prevalent in this population than in most other populations of asthmatics. However there is currently no data available to conclude that the prevalence of psychiatric disorders is lower in individuals with uncompensated OA or in individuals with work-exacerbated asthma. Further studies are needed to compare the correlation of psychiatric disorders and psychological distress with asthma specific quality of life measures, such as the AQLQ(S), in individuals with OA, work-exacerbated asthma and asthma that is unrelated to the workplace .
The AQLQ(S) emotional function subscale respondents report on different aspects that have been grouped in this domain by the developers of this questionnaire in which the items relate to three broad dimensions (concerns, anger and anxiety). When considering individual questions of this domain, the one about feeling afraid of getting out of breath was significantly associated with the PSI depression subdomain and PRIME-MD mood disorder whereas the questions about feeling concern about having asthma and about the need to use medication were significantly associated with PSI anxiety levels and PRIME-MD anxiety disorders respectively. Since our analysis was descriptive, we do not suggest to reduce items that have not shown significant correlation with psychologic distress or psychiatric disorders from the emotions subdomain of the AQLQ(S) questionnaire.
Whereas the PSI is a continuous measure that provides information about the number and severity of psychological symptoms, PRIME-MD diagnoses are categorical: individuals are classified as having a particular disorder or not based on having fulfilled a defined number of diagnostic criteria. Therefore, the severity of psychiatric disorders as measured by the PRIME-MD cannot be quantified . Clinically, anxiety and depressive symptoms (and disorders) overlap significantly, so it can sometimes be difficult to determine if these disorders are separate entities or different manifestations of the same disorder . Due to the limited number of individuals with OA included in our study, we were not able to examine associations between AQLQ(S) scores and individual mood and anxiety disorders (e.g. panic disorder, generalized anxiety disorder). To demonstrate these associations, studies with larger samples of individuals with OA are needed. Further our sample consisted of manly male workers with OA and therefore one must be cautious when extrapolating our findings to a population of female workers as the prevalence of the different forms of psychiatric disorders might be different .
Our study does not allow us to determine the relation of causation. We did not have information available about psychological distress or any psychiatric disorder prior to the development of OA or at the time the diagnosis of OA was made. Furthermore, we did not gather information about concurrent or past behavioural or medical therapy for psychiatric disorders in each individual. To our knowledge, the prevalence of psychiatric disorders at the time of diagnosis of OA and its devolution after removal from the causing agent or workplace is currently unknown and thus other studies are needed to investigate these factors in prospective investigations. It is unclear how interventions specifically targeted to decrease psychological distress or psychiatric disorders change the natural course of both conditions OA and concurrent mental disorders.
An important strength of the present study is that extensive objective assessments including spirometry, measurement of nonspecific bronchial hyperreactivity, and specific inhalation testing were performed in all individuals, the latter of which is considered the reference standard for a diagnosis of OA [44, 45].