In a recent study published in Scientific Reports, researchers investigated the relationship between pulmonary artery catheter (PAC) use and clinical outcomes in cardiac surgery patients, focusing on in-hospital deaths and hospital stay durations both generally and within specific subgroups.
Study: Association of pulmonary artery catheter with in-hospital outcomes after cardiac surgery in the United States: National Inpatient Sample 1999–2019. Image Credit: J-THE PHOTOHOLIC/Shutterstock.com
PACs provide specialized hemodynamic data essential for cardiac surgery decisions, yet while they may lead to more intensive treatments, improved outcomes are not always guaranteed.
Despite their potential utility, the absence of randomized trial data has led global guidelines to somewhat discourage PACs' routine use, highlighting the balance between its potential benefits and the scarcity of evidence in certain scenarios.
This stance is further complicated by inconsistent usage across nations, fueling the ongoing debate surrounding PACs.
PACs deliver direct and indirect hemodynamic measurements, assisting in treatment selection and monitoring chronic conditions. Some doubt their precision, especially concerning cardiac output (CO) estimations.
Limited studies have examined differences in outcomes, like in-hospital mortality and intensive care unit (ICU) admission, between PAC users and non-users. A knowledge gap persists regarding PAC's utility for certain patient subgroups. This highlights the need for further study investigating PAC's impact on cardiac surgery outcomes.
About the study
In the present study, analyses were conducted on data from the Agency for Healthcare Research and Quality (AHRQ) Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS) between 1999–2019, which holds hospital discharge records from the United States (US) community hospitals.
The data encompasses patient demographics, diagnoses, procedures, and hospital outcomes. The Fort Belvoir Community Hospital deemed the study exempt from human subject considerations.
The study focused on patients 18 years and older with a record indicating cardiac surgery. Records from hospitals with fewer than 50 cardiac surgeries from 1999 to 2019 and those with zero PACs yearly were excluded, after which 969,034 records remained.
Patients were categorized by age, sex, race/ethnicity, Charlson comorbidity index, elective admissions, admission quarter, weekend admission status, and primary payer. Comorbidities relevant to cardiovascular surgery patients were also taken into account. Hospitals were categorized by region, control, location, teaching status, and bed size.
Specific subgroups were identified using diagnostic and procedure codes: "heart failure," "pulmonary hypertension," "mitral or tricuspid valve disease," and "combined surgery."
Hospital discharge records were labeled based on whether the patient received a PAC or not. Outcomes considered were in-hospital mortality and length of stay (LOS).
Using Stata version 17, analyses were conducted incorporating descriptive statistics, bivariate associations, linear and logistic regression models, risk adjustment, and targeted maximum likelihood estimation (TMLE).
Sensitivity analyses were conducted based on hospital-level PAC rates. A significance level of P < 0.05 was adopted. The Fort Belvoir Community Hospital Institutional Review Board waived the study due to its non-human subject nature, adhering to the Declaration of Helsinki.
The study results reported that fewer disparities were noticed concerning PAC receipt between patients with selected subgroup characteristics than those without.
Among cardiac surgery patients, the PAC receipt rate was 9.49%. However, this was higher for patients with subgroup characteristics (10.01%) than those without (9.23%, significant at P=0.007).
The in-hospital mortality rate in the cardiac surgery group stood at 4.05%. Remarkably, there was a difference based on subgroup characteristics: 3.05% for patients without any and 6.06% for those with any of the characteristics (significant at P<0.0001).
The average hospital stay was 11.40 days, with patients having subgroup characteristics staying longer (13.37 days) than those without (10.41 days, significant difference at P<0.0001).
When comparing PAC recipients and non-recipients, there were noticeable variations in hospital stay lengths, but in-hospital death rates remained relatively consistent.
After accounting for various factors, the chances of in-hospital death did not show a significant difference between PAC recipients and non-recipients. Interestingly, in certain risk-adjusted models, PAC recipients had shorter hospital stays than non-recipients.
The results mirrored those from risk-adjusted models when looking at causal relationships between PAC receipt, in-hospital deaths, and hospital stays. PAC receipt was significantly associated with in-hospital death due to congestive heart failure patients. Moreover, PAC recipients were generally less likely to have hospital stays exceeding seven days.
In-depth stratified analyses based on subgroup status showed no major difference in the relationship between PAC and clinical outcomes. Sensitivity tests, when sorted by hospital PAC rates, showed distinct trends.
In particular, hospitals with PAC rates in the first to third quartiles had more in-hospital deaths, either no difference or longer stays for PAC recipients than non-recipients. In contrast, hospitals in the fourth quartile showed fewer deaths and shorter stays for PAC recipients.
Lastly, PAC use was not linked to in-hospital death upon excluding combined surgeries and focusing on specific cardiac surgeries. However, it was associated with shorter stays for mitral valve repairs and longer stays for tricuspid valve repairs.