A new formula accurately calculates the risk of pulmonary hypertension by predicting the mean pulmonary artery pressure using data obtained from pulmonary function testing and pulse oximetry, according to a report by Dr. Benjamin E. Schreiber, of the Royal Free Hampstead National Health Service Trust, London, and his associates.
They developed the formula based on findings in 257 patients with systemic sclerosis; the investigators validated it with data from another 129 similar patients.
Patients with a predicted mean pulmonary artery pressure (mPAP) below 25 mm Hg had a low pulmonary hypertension (PH) prevalence (4.4%). Those with a predicted mPAP between 25-35 mm Hg had an average PH prevalence (11.3%). Finally, those with a predicted mPAP above 35 mm Hg had a high PH prevalence (62.9%), which justifies using right-sided heart catheterization (RHC) (Arthritis Rheum. 2011;63:3531-9).
The study started with 838 patients who have systemic sclerosis (SSc) who had undergone an RHC. Of those, 386 patients had undergone pulmonary function testing (PFT) within 6 months of the RHC. Data on oxygen saturation as measured by pulse oximetry (SpO2) at the time of the RHC were included in the study. Of the total number of patients, 452 were excluded because they had incomplete PFT data.
The investigators randomly split the remaining 386 into two groups: the derivation and validation cohorts. The derivation cohort had 257 patients and its data were used to create the formula. The validation cohort had 129 subjects and its data were used to validate the formula. An additional 155 patients with connective tissue diseases other than SSc were obtained from the Royal Free Pulmonary Hypertension Service in London, United Kingdom.
The formula was derived using multivariable linear regression in the derivation cohort. Dr. Schreiber and his team examined the relationship between the SpO2, PFTs, clinical subtype, autoimmune serologic features, and the mPAP on RHC and developed the following formula: predicted mPAP = 136 – SpO2 –0.25 x diffusing capacity for carbon monoxide (DLCO)% predicted.
Applying the formula to the validation cohort gave an area under the curve (AUC) of 0.75. The formula performed best in patients with anticentromere antibodies, with an AUC of 0.87. In patients with connective tissue diseases other than SSc, the formula gave an AUC of 0.64.
The researchers suggested the PFT-derived formula can complement data obtained through echocardiography, and that both tests are predictive. Echocardiography is a good tool for diagnosing PH, but it cannot evaluate lung disease and is not always feasible to conduct. Therefore, the formula can calculate which patients are most likely to have PH.
Some of the limitations of this retrospective study included missing data, especially from echocardiograms and PFTs, and lack of rigorous selection of patients for RHC. There were no echocardiographic data available for most patients, and there may have been selection bias in the patients who had a second echocardiography performed.
Dr. Schreiber and his team suggested that further research should combine PFT data with other noninvasive information, such as symptoms, 6-minute walking distance, and brain natriuretic peptide levels. They also said that they plan to continue to validate their formula in patients recruited into the DETECT (Detection of Pulmonary Artery Hypertension in Systemic Sclerosis) study.
No relevant financial disclosures were reported.