Aortic Valve Surgery Outcomes Reporting and Implications for Persons Who Use Drugs With Infectious Endocarditis

(See the Major Article by Kimmel et al on pages 480–7.)

In their interrupted time series analysis of a sample of US inpatient hospitalizations of injection drug use (IDU) related infectious endocarditis (IE) (IDU-IE) and non-IDU IE before and after public reporting of aortic valve (AV) surgery outcomes was instituted in 2013, Kimmel SD et al’s current study published in this issue of Clinical Infectious Diseases identified some important findings that have relevant implications from a public health standpoint. In this evaluation, hospital admissions for IE, AV surgeries, and in-hospital mortality among persons hospitalized with IDU-IE were compared to those with non-IDU IE in a time period before (2010–2012) the national reporting system was instituted in 2013 to after this policy was enacted (August of 2013–2015). This study utilized International Classification of Diseases, 9th Revision (ICD-9) coding data from the National Inpatient Sample (NIS) database. The importance of undertaking this analysis is underscored by dramatic increases in IDU-related infections [1] in the setting of ever-worsening epidemics of opioid [2, 3] and non-opioid overdoses [4] in the United States. As highlighted by Kimmel et al, concern has been raised in the medical community that because of a rise in bacterial and fungal infective endocarditis diagnoses among persons who use drugs (PWUD) [5, 6], surgeons may be less inclined to perform surgery due to concern for ongoing drug use and risk of reinfection after valve surgery and other biases [7, 8]. Thus evaluating the potential negative impact of public reporting of IDU-IE on the ability to obtain life-saving valve surgery is important to understand.

In this study, first the authors showed that hospital admissions due to IE overall rose from 18.3% in the period before the national reporting occurring in 2013 to 27.3% after the institution of the national reporting period. These increases occurred among those who were younger, of white race, living in poorer neighborhoods, in receipt of Medicaid, and were admitted for nonelective reasons. These findings mirror national trends of rising infections among PWUD [1, 9–13] in the setting of the opioid and non-opioid drug epidemics in this country [2, 3, 9]. Although both IDU and non-IDU IE hospital admissions increased during these periods, the IDU-IE admissions were predominately composed of those who were young, white, female, and receiving Medicaid compared to prenational reporting time periods. This study highlights the importance of Medicaid expansion that occurred in 2014, a year after the national reporting was instituted. It is likely that without Medicaid expansion in 2014, the hospital admissions would have decreased. Data for this study focus on receipt of AV surgeries and in-hospital mortality; thus we can only speculate on mortality changes in those who did not seek medical care or who were not hospitalized.

The primary outcome of this study was to assess changes in native-AV surgical procedure rates from the prenational reporting period to the postnational reporting period. Although there was a reported increase in AV surgery rates in the prereporting period among those with non-IDU IE by 3% from 2010 to 2012, interestingly there was a 1% decrease in the prereporting phase of IDU-IE from 2010 to 2012. Both IDU IE and non-IDU IE AV surgeries decreased by 4% in the postreporting period, suggesting that there were no differences in access to AV surgeries based on IDU status as gleaned from this data. When the data were further evaluated comparing pre- to post- public reporting AV surgery outcome periods, both groups had similar decreases in rates of AV surgeries (33% decrease in the IDU-IE group and 31% non-IDU IE group). Importantly, however, women and those from poorer neighborhoods were less likely to obtain AV surgeries overall, whereas those with private insurance were more likely to obtain AV surgeries. Increases in opioid and other illicit drug use-related overdoses are occurring among women [14] along with associated increases in injection drug use- and sexually acquired new human immunodeficiency virus (HIV) and hepatitis C virus (HCV) infections among women who inject drugs [12, 13]. More research is therefore needed to evaluate sex and gender differences in opioid use disorder (OUD) and other substance use disorders (SUDs) treatment integrated with their associated infectious complications. Additionally, private insurance was associated with an increase in receipt of AV surgeries in this study, further identifying a possible disparity in selection for surgery where those who have private insurance may be selected for IE related AV surgery over those without insurance or those receiving Medicaid. It is important to note that housing status was not evaluated in this analysis, but unstable housing and homelessness has been shown to be highly associated with poor medical outcomes [15, 16] along with poverty and lack of insurance especially in persons who use drugs.

The secondary outcome of this study was in-hospital mortality. Both the IDU and non-IDU IE groups had a 16% reduction of in-hospital mortality from the pre- to post-public reporting time periods. In particular, the IDU-IE group were 21% less likely to experience in-hospital mortality throughout the study compared to those with non-IDU IE. Sensitivity analyses demonstrated that the mean mortality risk and illness severity scores decreased among those who received surgery, whereas the risk scores increased among those who did not receive surgery yet there was no change in-hospital mortality. It is difficult to decipher from this analysis if there was a change in mortality risk scores due to surgical case selection after the public reporting was instituted. Future studies should include a longer time frame analysis including after discharge from the hospital to the community as suggested by the authors.

A notable limitation of this study was that it restricted data for analysis from the NIS with a diagnosis of native AV IE based on ICD-9 coding. There are well-known limitations as acknowledged by the authors in including the use of ICD-coded diagnoses for identifying illicit drug use to denote IDU-IE [17]. This limitation in itself identifies a very important public health problem: there is no standardized mechanism for reporting infections related to IDU in the United States other than for HIV and in some states for HCV infection. As such there is no national database of infectious complications of illicit drug use to evaluate. Further, the Centers for Disease Control and Prevention national databases for opioid and other substance use overdose deaths are not updated in real-time; rather they are a few years behind. Therefore, it is difficult if not impossible to pinpoint and prevent new epidemics of drug overdoses and associated infectious complications before they happen as of yet in the United States.

This study also importantly highlights that the vast majority of infections related to IDU are bacterial and fungal complications related to needle associated pathogen entry [18–20]. Infectious disease epidemics are worsening dramatically related to the opioid and other illicit substance use epidemics in the United States [21], and these infections not only include endocarditis but also include skin and soft tissue infections, septic arthritis, and osteomyelitis [9, 22]. Such infections among PWUD are known to occur at higher rates than HIV and HCV, yet there is no effective way to track them in a national real-time manner. In addition, National Institutes of Health dollars have been for the most part directed toward opioid overdose treatment and HIV infectious complications of IDU but not toward bacterial and fungal infectious consequences of the opioid and poly-substance epidemics.

Although there were overall decreases in both IDU and non-IDU related IE AV surgeries and in hospital mortality rates after public reporting in this study, we do not know whether there were selection biases among those chosen for surgery compared to those who were not chosen to have surgery. Further, in order to reduce morbidity and mortality related to IDU associated IE, better screening and integrated prevention and treatment of SUD and associated infections needs to be done. In fact, most persons are not screened for OUD or other SUDs while hospitalized; thus IE related to SUDs is likely underreported. Hospitalizations for persons with infections related to drug use is a very reachable moment [23] to identify treatable opioid and other SUDs and offer life-saving interventions [20].

What we do know is that there are ever-worsening opioid and other SUD epidemics occurring across this country causing death from overdoses as well as a rise in mortality and morbidity from associated infectious disease epidemics [21, 24]. Further, these coalescing substance and infectious disease epidemics are harming persons from poorer communities, the homeless, the uninsured, and women. Medicaid expansion and improved funding sources to integrate antimicrobial treatment of infections related to drug use with prevention and treatment of OUD and other SUDs in hospital and other clinical settings that includes medication treatment for OUD (e.g., buprenorphine) and harm reduction services like clean needles and syringes can reduce the harms related to drug use-associated infections like endocarditis [21, 24].

Note

Potential conflicts of interest. S. A. S. has received grant funding from NIDA (grant number K02DA032322) in support of this work for ongoing scholarship research and mentorship activities regarding interaction of opioid use disorder and infectious disease. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.