This study demonstrates the difficulties in diagnosing obstructive airway diseases in general practice solely on the basis of clinical signs and symptoms. A combination of clinical signs and symptoms with CRP had the best diagnostic accuracy. CRP was related to smoking history; and it could be shown for the first time that CRP is related to the severity of dyspnea in patients with OAD. Also FEV1 and FIV1 were associated with the severity of dyspnea.
Only self reported wheezing was positively associated with the presence of the disease, but the other parameters of medical history were not accurate for identifying patients suffering from OAD. These findings are consistent with previous studies [18–20]. Interestingly, coughing is negatively associated with the presence of OAD. This might contradict common clinical findings , but these are often derived from selected clinical settings. Coughing also had a negative association in one survey, pointing out that this symptom is often attributed to other illnesses than asthma . The results of auscultation are similar to those from Strauss et al, who found a positive LR of 2.7 in his survey. Thus, auscultation could be a difficult marker when the patient is not symptomatic during the physical examination.
The relation between CRP and the severity of airway obstruction in COPD was illustrated before in a population based survey, which found an association between FEV1 and CRP . CRP has attracted more attention over the last years, as multiple associations with cardiovascular diseases , COPD , osteoporosis  and even depression  were identified. It is speculated that some of these diseases are caused by a low grade inflammation reflected by a small elevation of inflammatory markers , which might indicate a higher prevalence of systemic complications. Our findings also confirm that there is a chronic inflammatory process on a low level, since a no more than slightly elevated CRP (> 2 mg/l) has a significant diagnostic OR for OAD. This cut-off-point is in line with Sin et al.  who found their best threshold value for CRP at 2.2 mg/l. Our finding of a positive association between CRP and the number of pack years is supported by Gan et al., who detected that active smoking increases CRP . They suggested that smoking and reduced FEV1 have an additive effect on systemic inflammation. The association with the severity of dyspnea and smoking history might make CRP of interest for risk stratification of patients with OAD in primary care. The importance of this has been underlined by Huijnen et al., who identified dyspnea as a significant predictor for mortality . Until now, the diagnostic value of CRP for primary care was mostly evaluated for acute inflammatory diseases like lower respiratory tract infection (LRTI)  and acute maxillary sinusitis . In LRTI, CRP has a high diagnostic value in combination with clinical signs and symptoms . The impact on practice management was demonstrated as the use of CRP lowered antibiotic prescribing for sinusitis . In our study the diagnostic odds ratios of CRP improved in patients with dyspnea or smoking history. However, the confidence intervals were wide, which could mainly be due to the small sample size. And it must be noted additionally that it was not possible to differentiate between asthma and COPD. Therefore, further studies in larger populations of patients with COPD and asthma are necessary.
We were the first to demonstrate in primary care that dyspnea is correlated with FIV1. This relation was so far only demonstrated in a highly selected population with already known COPD [11, 12]. Taube et al found, that the perception of dyspnea is more related to the inspiratory than expiratory parameters . Thus, FIV1 could also provide possibilities for risk stratification of patients in primary care as it is related with the severity of dyspnea. However, it must be noted for our study, that FEV1 shows a higher correlation coefficient than FIV1. Therefore the applicability and specificity of these parameters for primary care need to be evaluated in further studies.
There are some more limitations of our study. A reason for underestimation of the prevalence of OAD could be that eight patients did not bring back the PEF-meter with the protocol. However, they assured that they had no problems any more. One could argue that the performance of bronchoprovocation in a specialized centre would have been the best gold standard to evaluate the accuracy of spirometry in primary care and to distinguish between COPD and asthma. As this was not available in general practice, we tried to get as close as possible to the theoretical gold standard . In a highly selected clinical population the measurement of PEF-variability within two weeks had a smaller sensitivity and specificity than bronchoprovocation . However, the value of bronchoprovocation should not be over-estimated as a poor agreement between bronchial hyper-reactivity and clinical asthma was demonstrated in a review , and the correlation between the clinical diagnosis asthma and bronchoprovocation could be low . Classical test characteristics derived from hospital studies are of limited value in primary care due to the lower incidence and smaller extent of the particular disease found there .