Pulmonary artery catheters (PAC) were introduced in the 1970s by professors Jeremy Swan and William Ganz [1]. Nowadays, the classical PAC has a nearly 50-year history of clinical applications for hemodynamic monitoring. In recent years, PAC have evolved from a device capable of combining static pressure measurements of intermittent cardiac output to a monitoring tool that could continuously measure cardiac output, oxygen supply, demand balance, and right ventricular function [2]. PAC utilization was valuable in guiding treatment in critically ill and complex surgical patients [3, 4, 5, 6], especially for patients undergoing cardiac surgery [7, 8, 9, 10]. Shaw et al. [7] reported that PAC utilization was associated with decreased length of stay (LOS), reduced cardiopulmonary morbidity, and increased infectious morbidity but no increase in 30-day in-hospital mortality during adult cardiac surgery. This suggested a potential benefit of PAC utilization in cardiac surgical patients [7].

PAC provided hemodynamic data, which were used to determine treatment measures. The decisive question was whether these treatment measures improve outcomes. For patients undergoing coronary artery bypass graft (CABG) surgery, PAC remains the most commonly used monitoring method among cardiovascular anesthesiologists [8]. For the studies that evaluated the effect of PAC utilization on the outcomes of CABG surgical patients, very few studies reported that the benefits of PAC outweigh the risks [9, 10]; two studies reported a neutral effect of PAC on clinical outcomes [11, 12]; several studies reported the harm of PAC [13, 14, 15, 16, 17]. Given the current use of PAC in off-pump coronary artery bypass (OPCAB) grafting surgery, the present study analyzed records of patients undergoing OPCAB surgery with or without PAC to estimate the impact of PAC use on short-term clinical outcomes.

Since its introduction in 1970 by Professor Jeremy Swan and William Ganz [1], PAC inserted at the bedside has been used in perioperative settings as a diagnostic tool and continuously monitors various physiological and hemodynamic parameters of critically ill surgical patients [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17]. However, the effect of PAC-guided goal-directed treatment on morbidity and mortality has not been clearly proven, and PAC utilization has been shown to increase healthcare service and hospitalization costs [4, 9, 13, 14, 17]. The SUPPORT study involved medical and surgical patients and showed PAC utilization had increased mortality, LOS in ICU, and hospitalization costs [4]. This study raised benefit-risk concerns over PAC utilization and generated intense interest in social medicals and professional circles. The PAC-man study enrolled 1041 ICU patients from the United Kingdom (UK) and showed no difference in hospital mortality between patients managed with or without PAC [6]. Subsequent consensus statements recommended redoubled efforts at education regarding the use of pulmonary-artery catheters. Thus, PAC use has declined in ICU and non-cardiac surgery settings. For example, one study reported that PAC use decreased by 65% from 5.66 per 1000 medical admissions in 1993 to 1.99 per 1000 in 2004 in surgical and critically ill patients [27]. Another survey reported PAC utilization in heart failure (HF) patients decreased by 67.8% from 6.28 per 1000 admissions in 1999 to 2.02 per 1000 admissions in 2013 [28]. However, other studies indicated that PAC was still highly used in cardiovascular surgery. For example, Judge and colleagues surveyed 6000 members of the Society of Cardiovascular Anesthesiologists (SCA) and showed that most respondents preferred using PAC for most cardiac patients [8]. Similarly, Brovman and colleagues reported that PAC utilization increased from 2010 to 2014 in cardiac surgical patients [29].

The present study analyzed records of patients undergoing OPCAB surgery with or without PAC to estimate the impact of PAC use on short-term clinical outcomes and found PAC utilization was associated with a higher likelihood of using epinephrine during intra- and postoperative settings and higher hospitalization costs. Still, it did not affect hospitalization mortality, and no significant intergroup differences in numerous adverse outcomes were noted, including stroke, neurological events, and renal replacement therapy. The effect of PAC utilization on long-term prognosis has also been reported. Xu and colleagues [17] reported that PAC use in CABG patients was neither associated with perioperative mortality and major complications, nor with long-term mortality and major adverse cardiac and cerebrovascular events. Particularly for renal replacement therapy, acute kidney injury was a strong predictor of clinical outcome after cardiac surgery [30]. Besides, PAC utilization also has limited benefits for patients undergoing coronary or valvular surgery. Brovman and colleagues reported that PAC utilization was not associated with improved operative mortality in the entire cohort of (n = 11,820), matched cohort (n = 7038), and in the recent heart failure, mitral valve disease, or tricuspid insufficiency subgroups. LOS in the ICU was longer, and there were more packed RBC transfusions in the PAC group; however, postoperative outcomes were similar, including stroke, sepsis, and new renal failure [29]. Thus, PAC should only be used when it is indicated.

It is worth noting that PAC is a user-dependent diagnostic and surveillance technique and not a treatment per se. Patients who underwent OPCAB surgery may benefit from timely-targeted and effective interventions guided by the comprehensive and real-time parameters provided by PAC. The anesthesiologists’ proficiency and experience with PAC were essential. Professor Jeremy Swan [31], the inventor of PAC, suggested that physicians place at least 50 PACs annually to maintain proficiency. Schwann and colleagues reported that PAC utilization during CABG was associated with increased end-organ failure rate and hospitalization death, possibly due to overtreatment with positive inotropic drugs and intravenous fluids in the PAC cohort [16]. This study also showed increased fluid infusion and epinephrine use in the PAC group. Tuman et al. [11] hypothesized that the PAC group was more likely to use drugs, partly reflecting how monitoring and unnecessary information influenced treatment without significantly altering outcomes. These results highlighted the risk of overtreating abnormal physiological parameters.

In 2003, the American Society of Anesthesiologists (ASA) updated the practice guidelines for PAC utilization, advising that the appropriate use of PAC should be based on three aspects, namely patient factors, surgery factors, and practice factors [19]. In 2021, the Chinese Society of Anesthesiology (CSA) recommended the indications of PAC utilization for cardiac surgical patients, including left ventricular systolic dysfunction (eject fraction 30%), right ventricular systolic dysfunction, left ventricular diastolic dysfunction, acute ventricular septal perforation and left ventricular assist device. In the present study, the administration of PAC was based on the anesthesiologist’s assessment. The anesthesiologist decides on PAC utilization or not, depending on various factors. Left ventricular function was the most commonly considered factor. Left ventricular dysfunction was associated with a high incidence of death and complications after cardiac surgery due to hemodynamic instability and potentially nonviable myocardium [32]. The key question was not whether PAC utilization improves outcomes but whether the treatment taken by the anesthesiologist based on PAC data improves outcomes. If a trained physician considers invasive hemodynamic data to be necessary to treat a particular patient, a PAC is justified.

Although the ideal evaluation of PAC in clinical practice is a randomized controlled trial (RCT), this effort is time-consuming, expensive, and has limited generalization [33].

We found that during the retrieval process, there was only one RCT study on PAC in CABG surgery, and it was published in 1989, with a small sample size (226) and selection bias [34]. In fact, most were observational cohort studies. This is the first and the largest survey to evaluate the relationships between PAC in OPCAB surgery and clinical outcomes in China. The results showed no differences in hospitalization outcomes between the OPCAB patients who used PAC and those who did not, except for significantly greater intra- and post-operative epinephrine use among the PAC group.

The current study demonstrated no apparent benefit or harm for PAC use during OPCAB surgery. However, PAC utilization in OPCAB surgery was more expensive.

The limitations of this study should be acknowledged. First, it is a retrospective study performed at a single center and using existing databases. The cohort could have been affected by selection bias. However, propensity score matching is used to overcome this discrepancy in the hopes of making each group more comparable regarding baseline risk. Fortunately, there are no significant differences in baseline characteristics between the PAC and no-PAC groups before propensity score-matching. Second, surgeons, anesthesiologists, and critical care physicians may have different levels of familiarity and expertise with PAC utilization, which may introduce some heterogeneity in the perioperative management of this population. Currently, there is no institutional protocol for managing hemodynamic and physiological parameters measured by PAC, which may lead to heterogeneity in perioperative care in this population. Furthermore, the effective use of PAC is an elusive variable that cannot be captured in this study. Moreover, as a single institution’s experience, all the subjects enrolled in the trial were from the same hospital in the same region, and the implications of these findings may not be fully generalizable.

The article’s data will be shared on reasonable request with the corresponding author.

CX: conceptualization, software, methodology, data collection/ curation, data analysis/interpretation, statistics, and writing-original draft. WQ, MS, LH: data collection, formal analysis, software, and critical article revision. YY: conceptualization, data collection, data analysis/interpretation, statistics, formal analysis, supervision, and critical article revision. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The Ethical Committee approved the single-center study (2019-1301). Because of the retrospective nature of this study, patient consent was waived.

The authors would like to thank the colleagues and statisticians at Fuwai Hospital and Fuwai Yunnan Hospital for their indispensable help with data collection and data analysis. We are also grateful to the CAMS Innovation Fund for Medical Sciences (CIFMS) 2021-I2M-C&T-B-038.

This research received no external funding.

The authors declare no conflict of interest.