Outcomes of noncardiac, nonobstetric surgery in patients with PAH: an international prospective survey

RESULTS

Of the 19 PH centres originally invited, 13 agreed to participate and 11 submitted their data (table S1). These 11 centres confirmed complete enrolment of all patients fulfilling the inclusion criteria. Thus, this survey enrolled 114 consecutive patients undergoing noncardiac, nonobstetric surgery from 11 PAH centres. Details of these patients are shown in table 1. At the time of surgery, all patients received targeted medical therapy for PAH and suffered from moderate-to-severe functional impairment, with 54% of them presenting in NYHA functional class I or II. The median 6MWD (assessed within 6 months before surgery) was 399 m. The last haemodynamic assessments prior to surgery (median interval between right heart catheterisation and surgery: 9 months, range 4–26 months) showed an average right atrial pressure of 7 mmHg, mean pulmonary artery pressure of 45 mmHg, pulmonary vascular resistance of 579 dyn·s·cm⁻⁵ and cardiac index of 2.9 L·min⁻¹·m⁻², indicating relatively well preserved right ventricular function in most patients.

The majority (82%; n=93) of the operations were performed under general anaesthesia, while the remaining 18% (n=21) were performed under spinal anaesthesia. 70 (61%) of the surgical interventions were elective, 31 (27%) were of intermediate urgency and 13 (11%) were classified as emergency interventions. 75% of all surgical procedures and 57% of the emergency interventions were performed in PH centres. There were no differences in baseline variables between patients undergoing emergency and nonemergency surgery (table S2). The majority of surgical interventions were abdominal, followed by trauma/orthopaedic surgery, gynaecological/urological surgery and others (table 1 and table S3).

The PAH medications were continued throughout the perioperative period in almost all cases. 20 (18%) of the patients had one or more PAH medications added during the perioperative period (inhaled nitric oxide n=4; inhaled iloprost n=12; i.v. prostacyclin analogues n=2; phosphodiesterase type 5 inhibitors n=3; and endothelin receptor antagonists n=1). Oral anticoagulants were always discontinued prior to surgery. Some centres routinely used pre-operative heparin when the international normalised ratio was <2.0 while other centres restarted anticoagulation only after surgery. Almost all patients were scheduled for intensive care unit treatment after the procedures. A Swan–Ganz catheter was used for perioperative haemodynamic monitoring in 15 (13%) patients.

Major complications occurred in seven (6.1%) of the surgical interventions (table 2). Three patients required catecholamine therapy for post-operative right heart failure and recovered. Four (3.5%) patients died. Three of the fatalities occurred after abdominal surgery and one was associated with hip replacement after traumatic fracture. Three of these interventions were performed under general anaesthesia and one under spinal anaesthesia. Causes of death were sepsis and/or right heart failure in all cases.

Risk factors for major complications after surgery were: a 6MWD <399 m and a right atrial pressure >7 mmHg on the last assessments prior to surgery; intra- or post-operative need for vasopressors; and the need for emergency surgery (table 3). The mortality rate associated with emergency surgery was two (15%) out of 13 compared to two (2%) out of 101 in nonemergency surgery (p=0.01 Fisher’s test).

DISCUSSION

To the best of our knowledge, this is the first prospective, multicentre assessment of outcomes after noncardiac, nonobstetric surgery in patients with a catheter-based diagnosis of PAH. Our data showed an overall mortality rate of 3.5%. The highest mortality rate (15%) was seen in emergency procedures, while the mortality rate in nonemergency surgery was 2%. Besides the need for emergency surgery, other factors associated with perioperative complications were lower pre-operative 6MWD, higher right atrial pressure and the requirement for catecholamines during the perioperative period.

The mortality rates observed in this survey appear somewhat lower than those reported in previous series of patients with PH undergoing surgery [8, 9, 10]. Comparisons, however, are difficult. Ramakrishna et al. [8] studied patients with various forms of PH and the diagnoses were mainly based on echocardiography (estimated right ventricular systolic pressure >35 mmHg) with confirmation by right heart catheterisation in only 68 (47%) out of 145 patients. Surgery-associated mortality was 7%, with the majority of deaths being attributed to respiratory or right heart failure, but it is unclear how many of these patients had an ascertained diagnosis of PAH. In contrast, Minai et al. [9] evaluated 28 surgical procedures in 21 patients with well-defined PAH and reported a mortality rate of 18%, predominantly due to right heart failure. Both series enrolled most of their patients before 2002; i.e. before most of the modern treatments for PAH became widely available. A more recent study from France reported two deaths (7%) among 28 patients with PAH or chronic thromboembolic PH enrolled between 2000 and 2007 [10]. Of note, the two fatal outcomes seen in that study occurred during emergency procedures.

The present series provides the largest data set of patients with a well-established diagnosis of PAH undergoing surgery and it indicates that the associated mortality, although still considerable, is not as high as previously reported, especially in nonemergency procedures. It is important to note that all of our patients received targeted therapy for PAH and their diseases appeared relatively well controlled, as indicated by a median 6MWD of 399 m at the last assessment (within 6 months before surgery) and the fact that 54% of the patients were in NYHA functional class I/II at the time of surgery. This may, in part, explain the low mortality rate of 2% among patients who had nonemergency surgical interventions. Specifically, only one of these patients died from right heart failure, while the other died from sepsis after cholecystectomy. The vast majority of patients who had elective surgical interventions survived, most of them without experiencing major complications. Elective procedures were preferentially performed in the PH centres, which might have contributed to the relatively good outcome. The sample size of the present study was too small to assess the effect of centre volume and experience on the post-operative outcome. We were unable to collect detailed data on the perioperative management of all patients, but it appears that modern anaesthesiology is capable of preventing haemodynamic deterioration in most cases. Although there is consensus amongst anaesthesiologists that regional anaesthesia is preferred over general anaesthesia in patients with PH [18, 19], there is no indication from our study that general anaesthesia is associated with more frequent or more serious complications than regional anaesthesia. However, our study was not sufficiently powered to compare the risks and complication of these approaches. Two recent papers on pregnancy in patients with PAH from two centres using disparate approaches to anaesthesia (one utilising almost exclusively epidural/spinal anaesthesia and one using exclusively general anaesthesia) reported similar outcomes, suggesting that both approaches are appropriate [20, 21].

In addition, the present study was not designed to assess the usefulness of intra-operative haemodynamic monitoring using a Swan–Ganz catheter. Since the majority (87%) of the surgical interventions did not use a Swan–Ganz catheter, it appears that they are not necessary in patients with stable disease. However, it is unclear whether the use of Swan–Ganz catheters or other devices to measure pulmonary vascular pressures and cardiac output may help to improve the outcome of selected patients presenting in an unstable haemodynamic situation.

There are several limitations to this study. Although being one of the largest surveys of its kind, the sample size is still small and the number of events is limited, prohibiting detailed analyses of risk factors for major complications and death. As this was a multicentre, multinational survey many centres used anaesthetic strategies; details of which we did not collect. Missing data cannot be fully excluded although all centres did follow their patients on a regular basis, making it unlikely that surgeries would have been missed, especially when associated with serious complications or even death. Our study included mainly patients with relatively well-controlled PAH, indicating that surgery is avoided in patients with advanced disease. Our results may therefore not be applicable to patients with severe PAH and/or manifest right heart failure. Importantly, there are no standard guidelines for management of these patients in the perioperative or post-operative periods. Our physicians may have applied simple measures such as meticulous use of anticoagulants, fluid balance, oxygen requirements, blood pressure or heart rate goals that allowed for better outcomes. Future studies are needed to determine this in detail.

In summary, noncardiac, nonobstetric surgery in patients with PAH is still associated with a considerable risk, but morbidity and mortality are lower than previously reported, at least in patients with well-controlled disease undergoing elective surgery. However, emergency procedures in patients with PAH continue to be associated with a high mortality.