Racial and ethnic disparities in antifibrotic therapy in idiopathic pulmonary fibrosis

Racial disparities have been documented in care of many respiratory diseases but little is known about the impact of race on the treatment of interstitial lung diseases. The purpose of this study was to determine how race and ethnicity influence treatment of idiopathic pulmonary fibrosis.


INTRODUCTION
Racial disparities have been shown to greatly influence outcomes in patients with various lung diseases. 1,2For instance, according to the United States Centers for Disease Control and Prevention, Black patients have a three-fold higher mortality for asthma compared to White individuals, and hospital readmission rates are significantly elevated for Black patients after acute asthma exacerbations. 3,4Moreover, studies have shown that Black individuals are more likely to develop chronic obstructive pulmonary disease (COPD) at an earlier age, have worse lung function decline over time, and experience higher mortality for various respiratory conditions, including COPD and lung cancer. 5,6Some studies have also reported a lower proportion of Black patients receiving optimal treatment for lung diseases, including asthma, COPD and lung cancer. 7Taken together, these findings highlight the significant impact of race on the care of patients with respiratory diseases.
While racial disparities have been thoroughly documented in many respiratory conditions, few, if any, investigations have explored the impact of race or ethnicity on patients with interstitial lung diseases (ILD). 7,8Existing studies suggest that Black patients are diagnosed with ILD at a younger age compared to White patients 9,10 and have a lower prevalence of idiopathic pulmonary fibrosis (IPF) and chronic hypersensitivity pneumonitis. 11One study suggested that Black patients have a better overall survival for non-IPF forms of ILD, whereas separate studies have shown that Black individuals have a higher mortality from IPF 12 but do better than White patients after acute exacerbations of the disease. 11,13While inconclusive, these studies provide support for the notion of race having an influence on diagnosis, management and outcomes in ILD.
In a recent quality improvement initiative to address racial disparities in our hospital, we uncovered that Black IPF patients were far less likely to be treated with antifibrotic agents (unpublished data) when compared to White and Hispanic patients.Indeed, we detected a nearly 40% decrease in the overall prevalence of antifibrotic treatments, Pirfenidone or Nintedanib, in Black versus White or Hispanic IPF patients.Together, these findings led to the hypothesis that race and/or ethnicity play an important role in the care of IPF patients.Thus, the main purpose of this study was to evaluate the impact of race and ethnicity on treatment for IPF in patients from multiple healthcare systems.

METHODS
Electronic health records were examined using TriNetX Research Network, a global health research network containing millions of patients.TriNetX contains de-identified data from more than 120 healthcare organizations and includes information on diagnoses, procedures, medications and laboratory values from patients in inpatient, outpatient and emergency room settings.While TriNetX includes countries from North America, South America, the European Union and Asia, the vast majority of data comes from institutions within the United States.However, our study restricted inclusion to IPF patients within US healthcare networks.Different institutions in the TriNetX network may join at different times, and their Electronic medical records are updated frequently.Most contributing health care organizations refresh data on average every 2-4 weeks, some daily.
TriNetX, LLC is certified to the ISO 27001:2013 standard and therefore ensures that healthcare data is protected and meets the requirements of the To generate cohorts, each group (White, Black, Hispanic and Asian patients) was paired with IPF diagnosis.Other racial groups, including American Indian or Alaskan Native and Native Hawaiian and other Pacific Islander, were excluded from analyses due to low sample size.Additionally, Sjogren syndrome with lung involvement (M35.02),systemic sclerosis with lung involvement (M34.81 and M34.0), idiopathic non-specific interstitial pneumonitis (NSIP) (J84.113),rheumatoid arthritis with lung involvement (M05.1),hypersensitivity pneumonitis (J67) and cryptogenic organizing pneumonia (J84.116) were explicitly excluded from each cohort to reduce the possibility of misdiagnosis within our IPF population.Race and ethnicity were self-identified.
All cohorts were matched for age, sex, history of nicotine dependence, dependence on supplemental oxygen, and predicted forced vital capacity (FVC) prior to performing comparisons (i.e., White patients with IPF vs. Black patients with IPF).Factors such as insufficient insurance and lowincome status were not used for matching purposes due to low percentage of each cohort having those characteristics.
A general IPF cohort was generated to analyse overall cohort characteristics, antifibrotic use and mortality at 3 years.Antifibrotic use for the general IPF cohort was calculated based on antifibrotic usage before and after index event of IPF diagnosis.Then, using our four IPF cohorts with respective race and ethnicity, we performed pair-wise cohort comparison of antifibrotic usage, specifically Pirfenidone or Nintedanib use, as the primary outcome.This was performed after propensity score matching through an algorithm created by TriNetX to generate matched subcohorts for respective analyses.Chi-square tests were used to compare outcomes between cohorts.

SUMMARY AT A GLANCE
We uncovered racial and ethnic disparities in the treatment of IPF.Black patients had lower antifibrotic use compared to White, Hispanic and Asian patients, even among older patients and those with lower lung function.White IPF patients on antifibrotics had a lower mortality, while treated and untreated Hispanic, Asian and Black IPF patients had similar mortality risk.
Initial comparisons were performed among all IPF patients but a secondary analysis was performed with IPF patients >65 years old, with the outcome being either Pirfenidone or Nintedanib usage.Since the prevalence of IPF is higher among older individuals, who also have a lower rate of autoimmune disease, we reasoned that IPF patients would be more accurately captured in this age-specific cohort. 13Additionally, we performed analyses on patients with lower lung function (predicted FVC of <3 L), with the outcome again being either Pirfenidone or Nintedanib usage, in order to determine whether treatment decisions were influenced by severity of disease.Like the larger cohort, these patients were matched for age, sex, history of nicotine dependence and supplemental oxygen use.However, due to small sample size, cohorts with less than 50 patients were excluded from our final analyses.Additionally, 3-year mortality was compared between IPF patients who were on versus not on antifibrotic therapy.Finally, we compared baseline patient characteristics among White and Black IPF patients as an attempt to identify factors that may have influenced antifibrotic treatment.Statistical analyses were performed using the TriNetX software version 3.4.4.

RESULTS
We identified a total of 47,184 patients with IPF from a total of 57 Health Care Organizations (HCO) in the TriNetX US Collaborative Network.Overall, the majority of IPF patients were White (35,082), followed by Hispanic (6079), Black (5245) and Asian (1221).This racial and ethnic breakdown of IPF patients parallels that from other studies. 14,15There were a majority of male patients (55%), and the average age of our total IPF cohort was 69.3 years old, which is similar to other IPF studies. 16,17Overall, the percentage of patients on Nintedanib versus Pirfenidone was similar among IPF patients and mortality at 3 years was 25.0%, as seen in Table 1.
After setting the age limit to over 65 years old, White patients had statistically greater antifibrotic usage compared to Black patients (18.5% vs. 9.5%, p-value <0.0001).Black individuals also had significantly less antifibrotic use compared to Hispanic (11.7% vs. 22.4%, p-value <0.0001) and Asian patients (13.0% vs. 18.9%, p-value = 0.003).Statistically significant differences in antifibrotic usage were not observed in other race or ethnic comparisons among those >65 years of age cohorts as seen in Table 3.
When defining cohorts as those with lower lung function (FVC of 0-3 L), only Black and White post-matching cohorts had greater than 50 patients, so other cohorts were excluded from these analyses.As shown in Table 4, we detected a significant difference in antifibrotic usage between White and Black patients with lower lung volume (33.8% vs. 18.3%, p-value = 0.04).
To determine the relationship between antifibrotic treatment and mortality we compared mortality in 3 years among IPF cohorts receiving or not receiving therapy.As shown in Figure 1 and Table 5, we detected a significantly lower three-year mortality for White patients on antifibrotic therapy (26.5% vs. 29.7%,p-value <0.0001).Mortality differences were not observed among Hispanic, Asian and Black individuals treated or not treated with antifibrotics.
Finally, to identify factors that may contribute to treatments disparities in our study, we evaluated differences in patient characteristics among White and Black patients.When compared to Black patients, we found that White individuals were significantly less likely to be incarcerated, unemployed, lack adequate food supply, be homeless, and have low income.White patients were also less likely to have specific comorbidities, such as asthma, pulmonary embolism, chronic passive liver congestion, hepatic failure and toxic liver disease as seen in Table S1 in the Supporting Information.

DISCUSSION
9][20] In this study, we show that Black IPF patients are less likely to receive antifibrotic therapy when compared to White, Hispanic and Asian patients.Further, this relationship persisted even in older patients, who have a higher probability of having IPF, and in patients with lower lung function, who presumably have advanced-stage disease.White patients on antifibrotics had a lower mortality compared to untreated patients, suggesting that antifibrotics alone or the care associated with providers who prescribe antifibrotics has a meaningful impact on IPF outcomes.Social factors can contribute to racial disparities in the treatment of many other diseases. 21,22Factors such as poverty and poor access to care and problems related to education and literacy have been linked to treatment differences in Black patients. 22,23Relevant to these reports, our analyses revealed that Black IPF patients were more likely to have problems related to homelessness, economic problems and imprisonment (Table S1 in the Supporting Information), which could have further limited their access to equitable healthcare treatment.Likewise, there is an aspect of cultural sensitivity that needs to be acknowledged in the context of treating Black patients, as physicians and their medical recommendations may be viewed with hesitance, potentially influenced by historical events like the Tuskegee study. 24In the United States, Black individuals have historically received less access to specialized care, and experienced bias and substandard care in their overall treatment. 24This is further compounded by historical practices outside of medicine that have resulted in further entrenched poverty, lack of education and employment and inequities that lead to large disparities in healthcare access, especially in the United States.Since the care of IPF patients often requires frequent follow-up, including visits with experts from multiple disciplines as well as various testing procedures, it would be expected that social factors like access to healthcare systems have a disproportionate impact on the care of IPF patients.Undoubtedly, this is compounded by the high cost associated with subspecialty medications such as antifibrotic agents.These results may be reflected in other conditions potentially due to differences in access to care, but this study first establishes empirical facts about IPF treatment to pave way for future more refined analysis.
Another aspect that may have contributed to treatment disparities in our study may relate to providers having limited experience caring for Black IPF patients.For instance, IPF is predominately a disease of White men, 2 and the ethnic/racial makeup of most IPF trials is disproportionality White. 25,26oreover, studies suggest that mortality for non-IPF ILD (other than Sarcoidosis) occurs at a younger age among Black individuals compared to White patients, and Black patients are less likely to die from acute exacerbation of IPF, 9,10 suggesting some providers may feel, rightly or wrongly, less of an urgent need to treat Black IPF patients. 9,27Finally, Black individuals had a high prevalence of liver disease in our study, which may have raised concerns among some providers since antifibrotics can have hepatotoxic effects.
Another interesting finding in our study was the observation that treatment with antifibrotics associated with a lower mortality in White but not in Black, Hispanic and Asian patients.Since Black, Hispanic and Asian patients are underrepresented in large clinical trials, it seems plausible that antifibrotics may work less effectively in these populations. 28e also acknowledge that findings may relate to the small size of these populations, suggesting our study may be underpowered to detect differences in these populations.Further, it is also possible that mortality rates differ among patients enrolled in clinical trials compared those in real-world general populations.Based on all of these factors, we recommend caution with interpretation of our mortality findings  given the retrospective nature of our investigation, and the lack of details related to other important aspects of patient care, and the experience of centres caring for IPF patients.Despite the large cohort analysed, we recognize our study has limitations.For instance, we recognize that our study is subject to selection bias, and misdiagnosis, similar to other large database studies and that our study results may have been affected by the failure to adjust for other conforming factors (such as certain drugs or comorbidities) Further, we found that antifibrotic usage was low in our IPF patients, which contrasts data from other studies. 29However, it should be noted that data from clinical trials are generally collected from hospitals with specialized expertise in IPF management, and these providers are likely to prescribe treatment at a greater rate than the general population.Furthermore, our database included institutions from urban and rural areas where access to specialized drugs may differ substantially.Unfortunately, the TriNetX platform does not permit researchers to parse out patients based on geographic location or associated health care organization.Finally, our study uncovered a much lower mortality in IPF patients at 3 years when compared to rates in the published literature.Indeed, others have reported rates of death or lung transplantation at 36 months much closer to 50%, 30 which again may point to a variety of factors (such as misdiagnosis).
In summary, our findings suggest significant racial disparities exist in the care of IPF patients and that Black individuals are less likely to be prescribed antifibrotics compared to White, Hispanic and Asian patients.If future studies validate these findings, it is crucial to investigate the factors contributing to these racial disparities in IPF.

F
I G U R E 1 Three-year mortality comparison among IPF patients who are on versus not on antifibrotics.**p-value <0.05 was considered significant.T A B L E 5 Three year mortality comparison among IPF patients on versus not on antifibrotics.
Characteristics of IPF population overall on TriNetX.
T A B L E 1 T A B L E 2 Matched cohort pair-wise comparison of antifibrotic use in IPF patients: All ages.Patients were matched for current age, sex, personal history of nicotine dependence, dependence on supplemental oxygen and predicted FVC.Statistical significance level: * P < 0.05, *** p < 0.001.Matched cohort pair-wise comparison of antifibrotic use in IPF patients: >65 years old.Antifibrotic use in IPF patients with predicted FVC of 0-3 L.
a T A B L E 3 a Patients were matched for current age, sex, personal history of nicotine dependence, dependence on supplemental oxygen and predicted FVC.Statistical significance level: ** p < 0.01, *** p < 0.001.T A B L E 4