Description of the condition

A hematoma, or a collection of blood inside the axilla or breast surgical site, is one potential complication following breast surgery. A hematoma can require additional surgery to drain and/or readmission to the hospital, and it may predispose patients to infection (Cheng 2011). Addressing these complications may limit the amount of money reimbursed to providers (Smith 2017), factoring into their decision to perform a lower yield procedure with a lower risk of complications (De Souza 2012; Smith 2017). Furthermore, hematomas may increase patients' emotional stress and physical pain (Kaoutzanis 2017). While the incidence of hematomas associated with breast surgery requiring return to the operating room is relatively low (less than 2.1%) (Collins 2012; Gobble 2014; Yan 2015), the large number of breast surgery procedures performed every year makes its negative consequences significant at a population level. Nonsteroidal anti‐inflammatory drugs (NSAIDs) are being increasingly used to treat postoperative pain, and it is unclear whether there is an association between perioperative NSAID use and an increased risk of hematoma development at the surgical site.

Breast surgery encompasses oncologic, reconstructive, and cosmetic procedures. An estimated 316,120 women were diagnosed with breast cancer in 2017 in the USA, with approximately 97% of stage I and II, 93% of stage III, and 31% of stage IV patients undergoing surgical treatment (ACS 2017). Commonly performed oncologic breast procedures include lumpectomy, mastectomy, sentinel lymph node biopsy, and axillary dissection. In 2017, members of the American Society for Plastic Surgery (ASPS) performed over 600,000 reconstructive and cosmetic breast cases, including implant‐based reconstruction, autologous flap reconstruction, mastopexy, and augmentation, among others (ASPS 2017). Approximately 29% of these were reconstructive procedures, while the remaining 71% were cosmetic procedures (ASPS 2017).

Description of the intervention

The American Society of Anesthesiologists (ASA) released its most recent practice guidelines for acute pain management in the perioperative period in 2016 (Chou 2016). Medication selection for perioperative pain management is guided by patient factors, but an underlying principle is a multimodal approach, that is, where two or more drugs with differing modes of action are used to treat acute surgical pain.

Opioid drugs remain a mainstay of analgesia; however, 20 years ago Kehlet 1997 introduced the now‐common recommendation of an "around‐the‐clock" regimen of a nonsteroidal anti‐inflammatory drug (NSAID, for example ketorolac, flurbiprofen, diclofenac, celecoxib) and/or acetaminophen (paracetamol), unless contraindicated. This concept has been more recently adapted into standardized breast surgical treatment plans (Batdorf 2015; Bonde 2015; Bonde 2016; Davidge 2013). NSAID use has demonstrated equivalent efficacy to opioids and similar postoperative bleeding when compared to controls in a wide range of surgical procedures (Gobble 2014). Perioperative NSAID use for patients undergoing endoscopic sinus surgery reduced postoperative rescue analgesics that included opioid use in many studies, with bleeding seen in 0.8% of patients (Svider 2018). Perioperative NSAID use in pediatric patients undergoing tonsillectomy concluded there was insufficient evidence to exclude an increased risk of bleeding (Lewis 2013).

How the intervention might work

NSAIDs inhibit cyclooxygenase (COX) enzymes, thereby reducing prostaglandin synthesis and an inflammatory response that causes pain. There are two types of COX enzymes: COX‐1 and COX‐2. Both types produce prostaglandins that promote inflammation, pain, and fever. Most NSAIDs are reversible inhibitors; however, aspirin binds permanently to COX enzymes, leading to a prolonged duration of effect.

The use of NSAIDs perioperatively may be associated with bleeding complications. This is because NSAID inhibition of COX‐1 reduces thromboxane A2, which mediates platelet aggregation. Most cells, including those in the stomach, express COX‐1, which provides a protective effect in gastric tissue, so NSAIDs' inhibition of COX‐1 enzymes can lead to bleeding from the stomach.

Non‐selective NSAIDs also inhibit COX‐2, and their effects can be different to those that inhibit COX‐1 enzymes. COX‐2 is the most important contributor to inflammation, hypertension, and possibly cancer. It is induced by immune cell factors, shear stress, and tumor promoters. Selective COX‐2 inhibitors target the inflammatory process while minimizing gastric and non‐gastric bleeding. They may reduce the risks of hematoma and other significant bleeding after breast surgery, while still providing adequate pain control in comparison to non‐selective COX‐1/COX‐2 inhibitors by reducing endothelial prostacyclin and consequently increasing platelet aggregation. In this regard, focusing on the NSAID ketorolac may be misleading, as this has the highest COX‐1 selectivity of all the NSAIDs (Cheng 2016; Jarupongprapa 2013; Schmidt 2016).

A retrospective analysis of perioperative ketorolac use in patients undergoing breast reduction surgery demonstrated a three‐fold increase in the likelihood of developing a hematoma and the need to return to the operating room for hematoma removal (Cawthorn 2012). A randomized controlled trial (RCT) comparing an NSAID (ketorolac) to a non‐NSAID (metamizol) for postoperative pain in elective plastic surgery reported postoperative bleeding in two patients receiving an NSAID that required a return to the operating room (Marin‐Bertolin 1997). Other studies have demonstrated no difference in bleeding rates between different NSAIDs (i.e. flurbiprofen or ketorolac) and placebo (Gobble 2014; Sun 2013). In patients receiving perioperative diclofenac, the risk of postoperative bleeding was higher than placebo; however, none of the patients needed reoperation for bleeding or hematoma (Cheng 2016).

A large number of studies have investigated whether NSAIDs are efficacious in reducing postoperative pain. Meta‐analyses of RCTs comparing opioid medication alone versus opioids plus NSAIDs in surgical patients found that the addition of ketorolac to intravenous morphine significantly improved pain scores and reduced analgesic use compared to intravenous morphine alone; however, these benefits were not as clear when intravenous patient‐controlled opioid analgesia was compared alone to a selective COX‐2 inhibitor or a nonselective NSAID (ASA 2012). Another meta‐analysis of RCTs found the addition of an NSAID to be superior to placebo in reducing postoperative pain (Gobble 2014). When selective COX‐2 inhibitors (rofecoxib, etoricoxib, celecoxib, and parecoxib) were administered preoperatively, they significantly reduced postoperative pain when compared to placebo (Nir 2016). However, COX‐2 inhibitors have been associated with early failure of vascular free flaps (i.e. tissue that is disconnected from its original blood supply and reconnected at another location in the body) due to thrombosis from a lack of thromboxane A2 inhibition and platelet aggregation (Al‐Sukhun 2006; Bonde 2017).

Recent evidence has demonstrated that perioperative NSAIDs (flurbiprofen axetil and ketorolac) may decrease the recurrence of breast cancer by inhibiting pro‐inflammatory and pro‐tumorigenic factors in patients undergoing surgery. These inflammatory factors impair the immune system and promote tumor recurrence and metastasis (Desmedt 2018; Forget 2013; Wen 2015). Other non‐NSAID pain medications lack these anti‐inflammatory characteristics. Even within the different classes of NSAIDs, results vary within patient demographics. A comparison of intraoperative ketorolac versus diclofenac in patients with an increased body mass index showed that ketorolac but not diclofenac reduced the incidence of distant metastasis (Desmedt 2018).

Why it is important to do this review

With the recent focus on the over‐prescribing of opioids in the literature, it is important to assess the effectiveness and safety of non‐opioid pain medication regimens including NSAIDs (HHS 2016). Clinicians have differing opinions on the safety of perioperative NSAIDs for breast surgery given the unclear risk/benefit ratio. A hematoma during a plastic surgery procedure can result in complete loss of the reconstruction (Mikhaylov 2018).

Previous systematic reviews, focused on evaluating NSAID analgesics in breast surgery, were limited in scope (Cheng 2016; Gobble 2014; Stephens 2015). Stephens 2015 conducted a systematic review to assess the association between perioperative use of ketorolac and the incidence of hematomas in 981 patients undergoing face and breast plastic surgery procedures. Although they found a two‐fold increase in the incidence of hematomas in patients who received ketorolac, this difference was not statistically significant (Stephens 2015). A single‐institution study by Sharma 2001 also did not find any significant association between the use of ketorolac and the incidence of hematomas in breast reconstruction procedures. The relatively small sample sizes included in these studies may potentially account for the lack of association demonstrated between perioperative NSAID use and hematoma development. Additionally, these results addressed very specific patient populations, so the conclusions may not be generalizable to patients undergoing alternate breast surgery procedures. Furthermore, some countries rarely use ketorolac as the perioperative NSAID. An evaluation of other NSAIDs more commonly used around the world needs to be completed.

As perioperative NSAID use has demonstrated a clinical decrease of recurrence of breast cancer, it needs to be determined if this clinical effect would outweigh contraindications to receiving NSAIDs in these patients. The American Society of Plastic Surgeons (ASPS) recommends that patients stop taking NSAIDs, including aspirin, prior to breast surgery (ASPS 2018). Research has not yet clarified the risks and benefits of all NSAIDs between cosmetic and reconstructive breast surgery patients. Assessing different proportions of COX‐1/COX‐2 inhibition and patient outcomes may provide a more uniform standard of care. There is currently no Cochrane Review on perioperative NSAID use for breast surgery. These results will help lead to more conclusive surgical practice guidelines for plastic surgeons encountering perioperative outcomes.

The impact of NSAID use is a very complex topic. Although the focus of this Cochrane Review is primarily on surgical site bleeding outcomes, there are a multitude of important perioperative outcomes, including venous thromboembolism, major adverse cardiac events, acute kidney injury, gastrointestinal ulceration, and potential cancer recurrence or remission. These outcomes increase with age (Wongrakpanich 2018). In addition, other comorbidities can increase the frequency of these outcomes (Harirforoosh 2013; Wongrakpanich 2018). COX‐1 and COX‐2 inhibitor selectivity is highly relevant when assessing the risk‐benefit balance in this regard.