{
  "abstract": "Introduction Interstitial lung diseases (ILDs) cause inflammation and scarring of lung parenchyma, impairing gas exchange. With >200 ILD sub-types, it is important to predict disease course and prognosis. A common complication of ILD is pulmonary artery hypertension (PH), linked to worse outcomes. However, PH symptoms overlap with those of ILD (breathlessness, fatigue) and are easily missed. PH probability can be assessed via echocardiography, but this requires time, expertise and cost. Therefore, screening of those most at risk of PH is valuable, though no official guidelines exist. Rahaghi et al. (Chest 2022;162.1:145-155) highlighted this as a limitation and suggested screening via the transfer factor for carbon monoxide (TLCO), a measure performed as routine ILD monitoring which reduces with scarring and can be linked to PH probability. Here I audit our department’s screening practises for PH regarding TLCO measures.Methods A preliminary 258 ILD patients from a total of 854 (collection ongoing), were analysed. Three linear models were fit with %TLCO, FVC/TLCO and TLCO z-score as dependent variables, PH diagnosis category (positive/negative/no investigation) as the independent variable of interest, accounting for sex (F:114,M:144), age (22-91,mean=66.8±0.8) and ILD subtype (17 factor). Three binomial generalised linear models were then compared with the three lung function variables as explanatory variables and PH diagnosis as the dependent variable (1/0) to assess which variable had the best predictive power.Results 120/258 patients were referred for echocardiogram, and 48 were positive for PH (intermediate-high probability). Patients investigated for PH showed a significantly lower %TLCO and TLCO z-score than those not investigated, but there was no significant difference between those PH positive and negative ( figure 1). Only those PH positive showed a significantly higher FVC/TLCO ratio than those not referred and there was no difference between those PH positive and negative. %TLCO had the best predictive power for PH diagnosis (ΔAIC=-4.4,X2=6.441,p<0.05).Conclusions Lung function measures significantly differed between patients referred for PH investigation vs not, but not between those PH positive vs negative, suggesting lung function measures inform referral decisions but may have limited power to predict PH diagnosis. %TLCO appears to have the best predictive power for PH diagnosis.Abstract P14 Figure 1a) TLCO percentage predicted, b) FVC/TLCO Ratio, c) TLCO z-score for patients with no PH investigation, a negative PH diagnosis and a positive PH diagnosis. Statistical significance shown is from Tukey’s post-hoc tests. Violin plots show the spread of data, white points the mean and boxplots the median and interquartile ranges",
  "authors": [
    {
      "affiliations": [
        "University Hospital Llandough, Cardiff, United Kingdom"
      ],
      "name": "Jennifer Crawley"
    },
    {
      "affiliations": [
        "University Hospital Llandough, Cardiff, United Kingdom"
      ],
      "name": "Corey Davies"
    },
    {
      "affiliations": [
        "University Hospital Llandough, Cardiff, United Kingdom"
      ],
      "name": "Lois Attewell"
    }
  ],
  "title": "P14 Auditing use of lung function test measures for screening of pulmonary hypertension in patients with interstitial lung disease",
  "uid": "ba4d678e-20c9-5265-9c11-ca00b6a60d50"
}