Preoperative management of patients with ILD involves risk stratification as well as medical optimization of their lung disease and comorbidities. Accurate risk prediction is imperative to making an informed risk/benefit analysis of the decision to proceed with surgery.

There are multiple well-established patient-related risk factors for poor post-surgical outcomes in the general population. In the seminal systematic review for the American College of Physicians by Smetana et al., advanced age, American Society of Anesthesiologists (ASA) class 2 or higher, functional dependence, chronic obstructive pulmonary disease, and congestive heart failure were all identified as risk factors for PPC [4]. In the past, increased unadjusted risk of PPC in older age groups was attributed to accumulation of comorbidities; more recently, it has been demonstrated to be an independent risk factor [11, 12]. Additional risk factors to take into consideration include smoking status, pulmonary hypertension (PH), and recent respiratory infection prior to surgery [13,14,15].

Studies of patient risk factors for worse post-surgical outcomes in ILD-specific populations most often focus on complications following lung cancer resection and surgical lung biopsy (SLB). These procedures are highlighted in ILD populations not only because they directly involve the lung, but also because patients with ILD are at increased risk of developing lung cancer and may require SLB for diagnosis of the underlying ILD. Previously described risk factors for acute lung injury and other PPCs in patients with ILD undergoing lung cancer resection include the extent of resection, male sex, history of disease exacerbation, preoperative steroid use, reduced preoperative forced vital capacity (FVC) and diffusion capacity of carbon monoxide (DLCO), comorbidity and a usual interstitial pneumonia (UIP) pattern on chest computed tomography (CT) scan [9, 10, 16, 17]. Previously described risk factors for poor outcomes following SLB include the need for supplemental oxygen therapy pre-operatively, comorbid PH, acute exacerbation at the time of biopsy, low DLCO, and male sex [18,19,20].

Risk prediction tools have been developed that incorporate preoperative risk factors to calculate risk of postoperative complications. Canet et al. performed a multicenter cohort study of over 2,000 patients; the investigators divided the study population into subsamples for model development and model validation in order to generate a PPC predictive index [3]. This process led to the development of the ARISCAT (Assess Respiratory Risk in Surgical Patients in Catalonia) risk index, which assigns a weighted point score using seven independent PPC risk factors including age, preoperative oxygen saturation on room air, history of acute respiratory infection in the past month, preoperative anemia, intrathoracic surgery, surgical duration, and emergency surgery to predict PPC risk. This risk score was later prospectively validated in a large cohort of nearly 6,000 surgical patients in Europe, and is one of the most commonly used PPC-related risk prediction tools in clinical practice [21]. Other PPC-related risk scores exist, however are more limited in their definitions of PPC. The Arozullah risk index for example predicts postoperative respiratory failure, defined as mechanical ventilation for more than 48 h after surgery or reintubation after postoperative extubation; the index was developed specifically in men undergoing non-cardiac surgeries [22].

Despite the comparatively higher risk of worse post-surgical outcomes, no evidence-based models of risk stratification exist that are specific to the ILD patient population [23]. While the ARISCAT index is a helpful jumping off point, it is likely an underestimation of the actual risk of PPC in an ILD patient population. It, similarly to the other risk prediction tools, does not incorporate data points that are likely to be available as a result of routine clinical care of these patients, such as FVC and DLCO obtained from pulmonary function testing [9]. Other risk factors to take into consideration include the patient’s supplemental oxygen requirements and the presence of PH, a common comorbid condition in patients with ILD. We recommend performing spirometry, DLCO, 6-minute walk test, and transthoracic echocardiogram prior to any planned procedure in a patient with advanced ILD, assuming that this testing has not already been performed in recent months as part of the patient’s routine clinical care, and that there have been no changes in the patient’s underlying disease status in the interim.

Screening for obstructive sleep apnea (OSA) is recommended in patients with ILD, as it is highly prevalent in both IPF and non-IPF disease [24]. A recent meta-analysis reported the prevalence of OSA to be 76.4% in patients with IPF [25]. When available, a sleep study performed within the preceding 12 months is generally considered acceptable provided that there have been no major changes in weight, symptoms, or comorbidities during that interval; older sleep studies are less reliable reflections of current physiologic status [26]. Although robust prospective data relating to OSA in ILD remain limited, identifying underlying OSA is important, as it may prompt targeted management strategies to reduce perioperative risk [27].

A number of recent studies have evaluated the prognostic value of blood biomarkers related to systemic inflammation and nutritional status pre-operatively. For example, the controlling nutritional status (CONUT) score, which is calculated using serum albumin, total cholesterol, and lymphocyte count, was demonstrated as an independent PPC and one-year mortality prognostic factor in patients with resectable non-small cell lung cancer [28]. The aforementioned Arozullah preoperative risk index incorporates albumin level less than 30 g/L and blood urea nitrogen level more than 30 mg/dL, amongst other factors [22]. In patients with ILD specifically, a high CRP level has been supported as an independent risk factor for acute exacerbations following non-pulmonary surgeries, while circulating serum sialylated carbohydrate antigen KL-6 and lactate dehydrogenase levels have been associated with disease exacerbation following lung cancer resection [5, 16, 29]. We do not advocate for systematic use of these blood tests for purposes of pre-surgical evaluation at this time, while acknowledging there may be case-by-case utility.

Assuming that risk-benefit analysis leads to a decision to proceed with surgery in shared decision-making with the patient, it is important to take advantage of opportunities to optimize a patient’s disease status and address any modifiable risk factors prior to proceeding.

One important modifiable risk factor involves smoking cessation. Interestingly, smokers who attempt to decrease cigarette consumption shortly prior to surgery may not significantly reduce the risk of developing PPCs, and in fact may be more likely to experience PPCs compared to those who continue with baseline smoking habits leading up to the procedure [30, 31]. Possible explanations for this observed phenomenon include selection bias versus transient increases in sputum production in the short term aftermath of smoking cessation. Smoking cessation for at least two or more months pre-surgically has been recommended for sputum volume to decrease [32]. Additionally, elective surgeries are best postponed until complete resolution of respiratory tract infections.

Careful medication reconciliation is also a critical part of a patient’s pre-operative assessment. Patients with ILD will frequently be prescribed immunosuppression and/or anti-fibrotic medications to manage their underlying disease. Immunosuppressing medications in general decrease a patient’s ability to fight infection should one occur in the perioperative setting. There is notably no evidence that perioperative antibiotic therapy prevents postoperative pneumonia either in the general population, or in patients who are taking immunosuppressing antirheumatic drugs.

In 2022 the American College of Rheumatology and American Association of Hip and Knee Surgeons published guidelines regarding the management of antirheumatic medications in patients undergoing elective hip or knee arthroplasty as a modifiable risk factor for infection; these guidelines are not specific to patients with ILD, and there is uncertain generalizability to other orthopedic or non-orthopedic surgeries (Table 1) [33]. The guidelines do highlight an underlying framework for attempting to balance the risk of flare of diverse manifestations of SARDs with the risk of infection for patients who are prescribed immunosuppressing medications; the risk of ILD flare in this context should be further studied. This is a particular challenge when patients are taking medications with long dosing intervals that are unable to be readily discontinued, most notably rituximab which is generally dosed every four to six months. The utility of measuring immunoglobulin levels at the time of surgery for patients who are taking rituximab requires further investigation. In addition to rituximab, patients with SARD-ILD may be prescribed mycophenolate, azathioprine, tocilizumab, tacrolimus, and Janus kinase (JAK) inhibitors for treatment of alveolitis.

Nonsteroidal anti-inflammatory medications such as ibuprofen and naproxen are also sometimes used to treat underlying rheumatic diseases; these medications require careful consideration given risks of kidney toxicity and bleeding peri-operatively.

In addition to antirheumatic drugs, patients with SARD-ILD may also concomitantly be prescribed glucocorticoids, which are known to adversely affect wound healing. As such, conversion to steroid-sparing regimens should be carefully considered prior to planned procedures based on patient-specific factors. In patients who will remain on glucocorticoids during the time of their procedure, the aforementioned 2022 guidelines recommend continuing these medications at their daily dose rather than increasing to “stress dosing.” [33] This recommendation is not applicable to patients who are receiving glucocorticoids for treatment of primary adrenal insufficiency or primary hypothalamic disease; patients who have been taking an equivalent of 20 milligrams of prednisone or higher for a duration of three weeks or longer may also experience suppression of the hypothalamic-pituitary axis. There is a theoretical concern for hypotension in patients who are receiving chronic glucocorticoid therapy. However, in a study of 432 RA patients who underwent total hip or knee arthroplasty, perioperative glucocorticoid administration in doses exceeding chronic home therapy was not associated with decreased risk of hypotension, but was associated with hyperglycemia and other early postoperative complications [34].

Anti-fibrotic medications including nintedanib and pirfenidone also pose a theoretical risk of impaired wound healing based upon their mechanisms of action, though a number of small studies have reported on their safety in the perioperative period [35,36,37]. Investigators have also reported on their efficacy in preventing acute exacerbation of ILD (AE-ILD) peri-operatively, specifically in patients with IPF, though no established guidelines exist which comment on peri operative dosing regimens [38, 39]. Further prospective investigation is needed to determine whether preventive drug administration is beneficial for patients with pulmonary fibrosis of various etiologies who are undergoing surgical procedures, including in situations when the medications are not previously prescribed for chronic use.

Given high rates of comorbid cardiovascular disease in both IPF and SARD-ILD patient populations, clinicians should have a low threshold for cardiology evaluation pre-surgically, as functional limitations of patients with ILD may adversely impact a clinician’s ability to assess cardiac symptoms [40]. In addition to coronary artery disease, PH is another important cardiovascular consideration; its associated risk on surgical outcomes varies depending upon PH etiology and clinical status. While clinicians may most readily associate ILD with World Health Organization (WHO) group 3 PH relating to chronic lung disease, patients with SARD-ILD due to underlying systemic sclerosis or mixed connective tissue disease, among other autoimmune diseases, are also at risk of developing pulmonary arterial hypertension (PAH, WHO group 1 PH). Patients with PH resulting in right heart dysfunction are at risk of hemodynamic instability and subsequent circulatory collapse following administration of anesthesia induction medications and the initiation of positive pressure ventilation [41]. Additionally, chronic PAH-targeted therapies such as prostanoids may increase the risk of bleeding [42]. Patients with ILD and comorbid PH will benefit from having surgery in a center with experienced PH providers [43].

Another clinical scenario may involve patients with SARD-ILD who have concomitant anti-phospholipid syndrome (APS); these patients may require bridging anticoagulation therapy to decrease the risk of deep venous thrombosis and pulmonary embolism in the perioperative setting [44]. Periods without anticoagulation should be kept to a minimum for patients with a history of thrombosis, as thrombosis may occur even despite the use of prophylactic medications. Catastrophic APS is a rare though markedly severe manifestation of APS that may occur in a patient undergoing surgical stress; maintaining vigilance with regards to change in blood counts suggestive of thrombotic microangiopathy or evidence of other organ involvement is critical [45].

Several intraoperative factors have been shown to impact the risk of adverse outcomes. As reflected by their inclusion in the ARISCAT risk index in addition to data from other studies, procedure location (intrathoracic versus peripheral), urgency (emergent versus non-emergent), and duration all contribute to PPC risk [3, 8, 29, 46]. In a cohort of 282 patients with IPF, longer surgical time was independently associated with acute respiratory worsening, with each additional 60 min conferring increased risk (RR 1.03; 95% CI 1.02 to 1.05) [47]. Accordingly, minimizing operative duration as clinically feasible is generally recommended; this will in turn decrease the amount of time a patient is receiving general anesthesia (GA) and mechanical ventilation.

Avoidance of general anesthesia (GA) in favor of neuraxial and regional anesthesia should be considered when this approach is safe and acceptable; this recommendation is informed in part by data from patients with other chronic lung disease [48]. Even SLB and laparoscopic abdominal surgery as examples may be safely performed by non-GA methods [49, 50]. The residual effects of GA as well as neuromuscular blocking agents may increase PPC risk through hypoventilation and atelectasis post operatively, which can be particularly concerning in patients with ILD who have limited pulmonary reserve. In a large prospective observational cohort study that demonstrated increased incidence of PPC in patients who underwent GA, the administration of neuromuscular blockade reversal agents did not decrease risk using either sugammadex or neostigmine [51].

When GA and mechanical ventilation are performed in patients with ILD, lung-protective ventilation strategies, including tidal volumes six to eight mL/kg of ideal body weight and positive end-expiratory pressure (PEEP), in the operating room are critical to reduce risk of ventilator-associated lung injury [52]. Patients with ILD, which is oftentimes characterized by bibasilar lung fibrosis, have less recruitable lung tissue; thus these patients may have less improvement in hypoxemia by recruitment maneuvers or stepwise increases in PEEP. There is a theoretical risk of pneumothorax in patients with ILD resulting from high PEEP due to the presence of non-recruitable lung areas, leading to overdistention of normal lung; high PEEP has also been associated with increased mortality [53, 54]. Increases in intrathoracic pressure by way of PEEP up-titration may also increase the risk of hypotension; this risk of hypotension caused by lung overdistention is of particular concern in patients with concomitant PH. Attention to changes in mean arterial pressure and lung compliance following ventilator adjustments is advised while maintaining goal plateau pressures </=30 cm of water. Minimizing use of supplemental oxygen therapy to maintain peripheral capillary oxygen saturation at 88% to 92% is also advised; this strategy serves to decrease both absorptive atelectasis through nitrogen washout from alveolar space and oxidative stress through minimizing the formation of reactive oxygen species [55, 56].

Importantly, the risks of absorptive atelectasis and reactive oxygen species should be carefully balanced with the risk of hypoxemia in patients with PH, as hypoxia will increase pulmonary vascular resistance and subsequently increase right heart strain. Another consideration for patients with PH may include preferential use of norepinephrine in place of high doses of phenylephrine, the use of which may lead to unopposed pulmonary vasoconstriction and similarly stress the right heart.

A further concern for patients with SARD-ILD resulting from RA who are undergoing GA is the question of cervical spine disease involvement. These patients will benefit from special attention to neck movements during endotracheal intubation, and cervical spine radiographs with flexion and extension positioning can be considered to assess for atlanto-axial instability. Bronchoscopic intubation should be considered to minimize neck manipulation. Providers should also consider other areas of joint pain and limitations to mobility which may impact positioning during surgery.

Judicious use of intravenous fluids to achieve euvolemia while minimizing risk of volume overload and pulmonary edema following termination of positive pressure ventilation is also recommended. Higher intraoperative fluid balance has been associated with postoperative acute exacerbation in patients with IPF [57].