Letters
Our patient population was largely positive for ER, but only 40% received hormonal therapy. Despite this, only 4% of patients experienced a breast cancer–related event. Not performing a sentinel node biopsy may avoid potential morbidity and should be reevaluated for patients 70 years of age or older with clinically negative nodes. Patients were unlikely to have treatment recommendations changed based on a sentinel node biopsy, and adjuvant therapy was less likely to be administered, regardless of nodal status. Patients in this subgroup were more likely to die of causes other than breast cancer, and not performing a sentinel node biopsy did not affect survival. Alice Chung, MD Alexandra Gangi, MD Farin Amersi, MD Xiao Zhang, PhD Armando Giuliano, MD Author Affiliations: Division of Surgical Oncology, Cedars-Sinai Medical Center, Los Angeles, California (Chung, Gangi, Amersi, Giuliano); Department of Surgery, Cedars-Sinai Medical Center, Los Angeles, California (Chung, Gangi, Amersi, Giuliano); Department of Statistics, Samuel Oschin Cancer Center, Cedars-Sinai Medical Center, Los Angeles (Zhang). Corresponding Author: Alice Chung, MD, Division of Surgical Oncology, Department of Surgery, Cedars-Sinai Medical Center, 310 N San Vicente Blvd, 3rd Floor, Los Angeles, CA 90048 ([email protected]). Published Online: May 27, 2015. doi:10.1001/jamasurg.2015.0647. Author Contributions: Drs Chung and Gangi had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Chung, Giuliano. Acquisition, analysis, or interpretation of data: Chung, Gangi, Amersi, Zhang. Drafting of the manuscript: Chung, Gangi, Zhang, Giuliano. Critical revision of the manuscript for important intellectual content: Chung, Amersi, Giuliano. Statistical analysis: Zhang. Obtained funding: Gangi, Giuliano. Administrative, technical, or material support: Chung, Amersi, Giuliano. Conflict of Interest Disclosures: None reported. Funding/Support: The collection, management, and analysis of data for this work was supported by the Margie and Robert E. Petersen Foundation and the Fashion Footwear Charitable Foundation of New York. Role of the Funder/Sponsor: The funders/sponsors had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. 1. Jones EL, Leak A, Muss HB. Adjuvant therapy of breast cancer in women 70 years of age and older: tough decisions, high stakes. Oncology (Williston Park). 2012;26(9):793-801. 2. Rudenstam CM, Zahrieh D, Forbes JF, et al; International Breast Cancer Study Group. Randomized trial comparing axillary clearance versus no axillary clearance in older patients with breast cancer: first results of International Breast Cancer Study Group Trial 10-93. J Clin Oncol. 2006;24(3):337-344. 3. Martelli G, Boracchi P, De Palo M, et al. A randomized trial comparing axillary dissection to no axillary dissection in older patients with T1N0 breast cancer: results after 5 years of follow-up. Ann Surg. 2005;242(1):1-6. 4. Fisher B, Jeong JH, Anderson S, Bryant J, Fisher ER, Wolmark N. Twenty-five-year follow-up of a randomized trial comparing radical mastectomy, total mastectomy, and total mastectomy followed by irradiation. N Engl J Med. 2002;347(8):567-575. 5. Veronesi U, Orecchia R, Zurrida S, et al. Avoiding axillary dissection in breast cancer surgery: a randomized trial to assess the role of axillary radiotherapy. Ann Oncol. 2005;16(3):383-388.
684
6. Hind D, Wyld L, Beverley CB, Reed MW. Surgery versus primary endocrine therapy for operable primary breast cancer in elderly women (70 years plus). Cochrane Database Syst Rev. 2006;(1):CD004272.
COMMENT & RESPONSE
Selective vs Nonselective Nonsteroidal Anti-inflammatory Drugs and Anastomotic Leakage After Colorectal Surgery To the Editor We read with interest the article by Hakkarainen et al1 and wish to commend the authors on their attempt to shed light on the challenging question of nonsteroidal antiinflammatory drugs (NSAIDs) and the risk for anastomotic leakage after colorectal surgery. A growing body of evidence has implicated NSAIDs, particularly selective NSAIDs, as a risk factor for anastomotic leakage. These types of NSAIDs work by inhibiting the cyclooxygenase (COX) class of enzymes: COX1, which is present throughout the body (including the vascular endothelium, stomach, and kidneys) and COX2, which is predominantly found at the site of injury (triggered by inflammatory mediators).2 Hakkarainen et al1 found a 24% increase in the leak rate in their cohort of patients who underwent bariatric surgery, elective colorectal surgery, or nonelective colorectal surgery. In their subgroup analysis, the association was limited to nonelective colorectal surgery (odds ratio [OR], 1.70 [P = .01]), not elective colorectal surgery (OR, 1.13 [P = .36]). They acknowledged that they could not specify the NSAIDs used in their study,1 but they believed that nonselective NSAIDs, such as ketorolac tromethamine, were predominantly used. We feel that there are 2 points requiring emphasis and clarification. First, the classification of selective vs nonselective NSAIDs is often misunderstood. The term selective refers to the newer class of COX2 inhibitors, whereas nonselective NSAIDs are generally thought to indiscriminately inhibit both isoforms. In point of fact, however, this is not the case. Nonselective NSAIDs can inhibit either COX1 or COX2 preferentially. Diclofenac, a nonselective NSAID, behaves like a selective NSAID and, in fact, has an inhibitory profile similar to that of the selective NSAID celocoxib, preferentially inhibiting COX2.2,3 Ketorolac, another nonselective NSAID, preferentially inhibits COX1 more so than any other commercially available NSAID.2 Second, virtually all NSAIDs have the potential to inhibit both COX enzymes regardless of their selectivity. For example, celocoxib will inhibit 60% of COX1 enzymes at 80% of its maximal inhibitory concentration for COX2.2 In our article4 looking at the perioperative use of ketorolac and anastomotic leakage after elective colorectal surgery, we found no association with leakage (OR, 1.21 [P = .66]). We attributed this to ketorolac’s COX1 selectiveness. However, in our subgroup analysis, higher doses were associated with leakage (OR, 1.29 [P = .048]) for every 15-mg increase in ketorolac administered. Perioperative use of NSAIDs and anastomotic leakage remain important issues in the fast-track era of multimodal pain management. Future studies should address the selectivity of
JAMA Surgery July 2015 Volume 150, Number 7 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a Tulane University User on 10/12/2016
jamasurgery.com
Letters
the NSAIDs used and the doses administered over the perioperative period to effectively answer this question. Fady Saleh, MD, MPH Allan Okrainec, MD, MHPE Author Affiliations: Division of General Surgery, University Health Network, Toronto, Ontario, Canada. Corresponding Author: Allan Okrainec, MD, MHPE, Division of General Surgery, University Health Network, 399 Bathurst St, 8-MP 325A, Toronto, ON M5T 2S8, Canada ([email protected]). Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0635. Conflict of Interest Disclosures: None reported. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228. 2. Warner TD, Giuliano F, Vojnovic I, Bukasa A, Mitchell JA, Vane JR. Nonsteroid drug selectivities for cyclo-oxygenase-1 rather than cyclo-oxygenase-2 are associated with human gastrointestinal toxicity: a full in vitro analysis. Proc Natl Acad Sci U S A. 1999;96(13):7563-7568. 3. Klein M, Gögenur I, Rosenberg J. Postoperative use of non-steroidal anti-inflammatory drugs in patients with anastomotic leakage requiring reoperation after colorectal resection: cohort study based on prospective data. BMJ. 2012;345:e6166. 4. Saleh F, Jackson TD, Ambrosini L, et al. Perioperative nonselective non-steroidal anti-inflammatory drugs are not associated with anastomotic leakage after colorectal surgery. J Gastrointest Surg. 2014;18(8):1398-1404.
To the Editor We read with great interest the article by Hakkarainen et al1 published in JAMA Surgery, the results of which are intriguing. The authors report the results of a retrospective cohort study of patients undergoing bariatric or colorectal surgery using data from the Surgical Care and Outcomes Assessment Program linked to the Washington State Comprehensive Abstract Reporting System.1 Although we recognize the importance of this work and applaud the tremendous effort it represents, we would like to draw attention to a number of specific concerns regarding this study.1 An obvious limitation is that the dose, duration, and formulation of the nonsteroidal anti-inflammatory drugs (NSAIDs) are unknown, as are the patient histories of preoperative use of NSAIDs. Although this point is discussed by the authors, we would respectfully question the generalizability of the authors’ assumptions that the “likely” use of intravenous ketorolac tromethamine and caldolor adds support to previous findings showing an increased risk of anastomotic leak with the use of nonselective NSAIDs.2 Another concern is the younger age of the patients receiving NSAIDs because they would be more likely to have a different colorectal pathology than the older patients, and this may also have influenced the results. We also find the rather flexible approach to designs, definitions, outcomes, and analytical modes somewhat worrisome because this type of approach has been implicated in higher rates of “positive findings.”3 We also question why Hakkarainen et al1 felt the need to include heterogeneous groups of patients (ie, patients who underwent bariatric, elective, or nonelective colorectal surgery) in 1 group. We contend that this led to unnecessary subgroup analyses when dealing with colorectal procedures; an indi-
vidual group would have simplified matters. In this context, the wider issue of using large data sets within scientific research and the inherent flaws that can flourish as a result need to be considered.4 Hakkarainen et al1 conclude that the postoperative use of NSAIDs was associated with a significantly increased risk for anastomotic complications among patients undergoing nonelective resection. However, we urge caution and believe that the evidence presented in this study1 does not allow for causation to be established and, therefore, does not convince us to change our current practice of using individualized analgesic regimens. Noel P. Lynch, MCh, MB, Med Sci Emily Boyle, MD Eamon G. Kavanagh, MD, FRCSI Author Affiliations: Department of Surgery, University Hospital Limerick, Limerick, Ireland. Corresponding Author: Noel P. Lynch, MCh, MB, Med Sci, Department of Surgery, University Hospital Limerick, Dooradoyle, Limerick, Ireland (noellynch @rcsi.ie). Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0641. Conflict of Interest Disclosures: None reported. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228. 2. Gorissen KJ, Benning D, Berghmans T, et al. Risk of anastomotic leakage with non-steroidal anti-inflammatory drugs in colorectal surgery. Br J Surg. 2012;99 (5):721-727. 3. Ioannidis JP. Why most published research findings are false. PLoS Med. 2005; 2(8):e124. 4. The PLoS Medicine Editors. Why bigger is not yet better: the problems with huge datasets. PLoS Med. 2005;2(2):e55.
To the Editor I would like to make several comments in relation to the article by Hakkarainen et al1 recently published in JAMA Surgery. Being part of a hospital trust that performs large numbers of colorectal surgical procedures, we have a specific interest in the outcomes of this study.1 The benefits of nonsteroidal anti-inflammatory drugs (NSAIDs) with regard to the opioid-sparing effect cannot be denied. However, some surgeons are concerned about the risk of anastomotic leaks; therefore, they prefer not to use NSAIDs, despite the evidence of leakage being weak and the benefits of NSAID use being very clear. The authors1 acknowledge several limitations, but these do not diminish the use of NSAIDs in any way. When researchers are examining an outcome due to a particular drug, it is of paramount importance that the type of drug (ie, the name of the NSAID), the subtype (cyclooxygenase 1 vs cyclooxygenase 2), the dose, and the duration of therapy are known. Surely, at a clinical level, an anastomotic leak 85 days after surgery cannot be blamed on 2 to 3 doses of NSAIDs given postoperatively? Without any further breakdown of the data on the NSAID therapy that was used, no actual clinical link between the 2 events can be determined. Hakkarainen et al1 actually found a “significant” association between patients with low albumin levels and anastomotic leakage. Few surgeons would argue that patients who
jamasurgery.com
(Reprinted) JAMA Surgery July 2015 Volume 150, Number 7
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a Tulane University User on 10/12/2016
685
Letters
are hypoalbuminemic for whatever reason are at risk of wound breakdown. Are these the same patients who happened to have received NSAIDs? If any conclusion can be drawn from this study,1 it is that, for patients undergoing elective colorectal surgery, NSAIDs appear to be safe and to play an important role in enhanced recovery protocols. We do not agree that there is enough proof to indicate that NSAID use results in anastomotic leakage in patients who have undergone emergency colorectal surgery, but we do agree that NSAIDs should only be used if no other contraindications (such as general debility, renal dysfunction, and dehydration) are indicated. In conclusion, it is clear that proper prospective trials are required to finally answer this question because this study1 poses more questions than it answers. Kevin Doody, BMBS Margaret Coleman, FFARCSI, MSc, SEM Author Affiliations: Department of Anaesthesia, University Hospital Limerick, Limerick, Ireland. Corresponding Author: Kevin Doody, BMBS, Department of Anaesthesia, University Hospital Limerick, Limerick, Ireland ([email protected]). Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0644. Conflict of Interest Disclosures: None reported. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228.
Although a key safety measure, concern over anastomotic leak rates may mask an overall beneficial effect of NSAIDs’ anti-inflammatory and opioid-sparing properties on patient outcomes. Our prospective, multicenter evaluation of NSAID safety in 1500 patients found that administration within 72 hours of surgery was associated with a reduction in overall complications.4 In line with the findings of a previous multicenter study, we found no evidence of an association between mixed NSAIDs (predominantly ibuprofen) and anastomotic leak.5 The absolute risk difference in anastomotic leak between NSAID and non-NSAID groups in Hakkarainen and colleagues’ study was 0.6%. The clinical significance of this small effect size set against the potential benefits of NSAIDs remains to be determined. Three multicenter observational studies have demonstrated the safety of NSAIDs and, particularly, ibuprofen in elective colorectal patients. The existing evidence base is unlikely to be enhanced by further observational studies. A large randomized clinical trial is required to test the efficacy and safety of NSAIDs in reducing complication rates following elective colorectal surgery. Thomas M. Drake, BMedSci Dmitri Nepogodiev, MBChB Henry A. Claireaux, BSc(Hons) Author Affiliations: Student Audit and Research in Surgery (STARSurg) Steering Group, Birmingham, England. Corresponding Author: Dmitri Nepogodiev, MBChB, Academic Department of Surgery, Old Queen Elizabeth Hospital, Room 29, 4th Floor, Edgbaston, Birmingham B15 2TH, England ([email protected]).
To the Editor We commend Hakkarainen and colleagues1 for their contribution to the ongoing discussion regarding the association of nonsteroidal anti-inflammatory drugs (NSAIDs) and anastomotic leak following gastrointestinal surgery. A previous meta-analysis found NSAIDs are significantly associated with an increased risk for anastomotic leak.2 To our knowledge, the study by Hakkarainen et al is the first to suggest that this effect is limited to emergency colorectal patients only. These findings are based on a well-conducted analysis of clearly defined groups from a large prospective database. However, the patients in the NSAID group were significantly younger, with fewer comorbidities than the control arm, which makes it difficult to generate reliable effect estimates between treatment groups. Comparison between groups is further hindered by a lack of data on perioperative and postoperative variables known to be associated with anastomotic leak.3 Nonetheless, it would be interesting to perform a casematched analysis of the data from Hakkarainen et al and use factors associated with anastomotic leakage to generate improved effect estimates. The potency and adverse effect profiles of individual NSAID agents vary greatly. Data regarding choice of NSAID agent was unavailable in the study by Hakkarainen and colleagues. If correct, their assumption that most patients received ketorolac would limit the generalizability of this study to centers outside of the United States, many of which use ibuprofen in preference to ketorolac. 686
Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0806. Conflict of Interest Disclosures: None reported. Additional Information: The STARSurg Collaborative Steering Group members include Thomas M. Drake, Dmitri Nepogodiev, Henry Claireaux, Midhun Mohan, Chetan Khatri, Chia Kong, Lisa McNamee, Michael Bath, Stephen J. Chapman, James Glasbey, Ewen M. Harrison, Aneel Bhangu, and J. Edward Fitzgerald. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228. 2. Bhangu A, Singh P, Fitzgerald JE, Slesser A, Tekkis P. Postoperative nonsteroidal anti-inflammatory drugs and risk of anastomotic leak: meta-analysis of clinical and experimental studies. World J Surg. 2014;38(9): 2247-2257. 3. McDermott FD, Heeney A, Kelly ME, Steele RJ, Carlson GL, Winter DC. Systematic review of preoperative, intraoperative and postoperative risk factors for colorectal anastomotic leaks. Br J Surg. 2015;102(5):462-479. 4. STARSurg Collaborative. Impact of postoperative non-steroidal anti-inflammatory drugs on adverse events after gastrointestinal surgery. Br J Surg. 2014;101(11):1413-1423. 5. Klein M, Gögenur I, Rosenberg J. Postoperative use of non-steroidal anti-inflammatory drugs in patients with anastomotic leakage requiring reoperation after colorectal resection: cohort study based on prospective data. BMJ. 2012;345:e6166.
In Reply We are pleased that our study1 has generated debate and discussion regarding the perioperative use of nonsteroidal anti-inflammatory drugs (NSAIDs) among patients undergoing gastrointestinal surgery, and we sincerely thank the several teams of physicians and surgeons who have taken the time to respond. First, we would like to reiterate that we do not sug-
JAMA Surgery July 2015 Volume 150, Number 7 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a Tulane University User on 10/12/2016
jamasurgery.com
Letters
gest that our data firmly establish causality between perioperative use of NSAIDs and increased rates of anastomotic leak. We believe that our data are strong enough to warrant increased caution when deciding whether or not to use NSAIDs for patients undergoing emergency colorectal procedures. Nor do we dispute the possible benefits of NSAID use in enhanced recovery protocols for their opioid-sparing potential. It is important to note that our data and conclusions are focused on patients undergoing nonelective colorectal procedures, and thus the authors who defend the use of NSAIDs in recovery protocols may be reading too much into our conclusions and taking them out of context. We agree with all the authors who responded that further studies are warranted and needed to address this question and to analyze the type of drug used, the dosage, and the duration of therapy. In our study,1 we simply urge caution. Regarding the selectivity of various NSAIDs, we agree with those authors who commented that all NSAIDs are selective for either cyclooxygenase 1 (COX1) or COX2 to varying degrees at different doses. Both ketorolac tromethamine and ibuprofen are relatively nonselective at the doses typically given, with only a mildly increased inhibition of COX1 vs COX2, and we consider them nonselective. Despite the lack of specific drug information in the Surgical Care and Outcomes Assessment Program (SCOAP) registry, we are quite confident that ketorolac and/or ibuprofen represent the large majority of NSAIDs used. We can be this confident because of the collaborative nature of the SCOAP sites and because of our ability to directly query surgeons at participating sites about formularies and practice patterns. As to our initial inclusion of patients undergoing bariatric or elective colorectal procedures, this is a function of our initial interest in the potential effects of NSAIDs on gastrointestinal anastomotic healing. Not including data from these 2 patient populations in our study 1 could be construed as disingenuous because it would be different from the original research question that we set out to answer, and, obviously, the data from these 2 patient populations yield different results from analysis of the patients undergoing nonelective colorectal surgery. Finally, an error regarding the sample size referenced in the title of our Table 3 occurred during the proofing process. The sample size of the cohort of patients undergoing nonelective colorectal procedures was 1621 patients, not 9624. This error will be corrected online. Timo W. Hakkarainen, MD, MS David R. Flum, MD, MPH Author Affiliations: Department of Surgery, University of Washington Medical Center, Seattle. Corresponding Author: Timo W. Hakkarainen, MD, MS, Department of Surgery, University of Washington Medical Center, 1959 NE Pacific St, PO Box 356410, Seattle, WA 98105-6410 ([email protected]). Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0638. Conflict of Interest Disclosures: None reported. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228.
Helping Meet Surgical Needs in Under-resourced Settings: The Role of Task Shifting To the Editor We commend the timely article by Kotagal and Horvath1 that discusses the challenges of providing surgical care in under-resourced settings, and we look forward to more articles that address this important issue. As the authors point out, low- and middle-income countries have complex health systems with major discrepancies in quality of and access to care even within each country/region. In many low- and middle-income countries, private medical care contributes the lion’s share (80%) of the services provided, putting it out of the reach of millions.2 In a middleincome country like India, the overall physician to population ratio is 1:1800 (below 1:1000 established by the World Health Organization), while the situation in the rural areas of India is dire with only 25% of trained physicians available for 60% to 70% of the population. The Indian government reports that 70% of specialist (surgeons, pediatricians, and gynecologists) positions in rural government health centers are vacant. Proposals by governments to make rural service a mandatory component of medical education have been met with resistance by physicians and their associations. Surgeons and other specialists are reluctant to work in rural areas, pointing out a lack of infrastructure and opportunities for professional growth, among other reasons. As a way to break this stalemate, low- and middleincome countries have started programs that are geared toward training midlevel health care professionals to serve in rural and underserved areas. While these programs have traditionally been focused on managing medical conditions, training these health care professionals in basic obstetric and surgical care would be hugely advantageous in reducing preventable mortality and morbidity. The training should include recognition of signs and symptoms of common conditions needing surgical consultation, as well as management of simple surgical and obstetric emergencies. Task shifting and creating a cadre of skilled midlevel health care professionals for basic surgical care have been validated in studies3-5 from under-resourced settings that have shown that the outcomes associated with these programs are comparable to the outcomes associated with the care provided by physicians. The potential limitations of task shifting need to be addressed in a diligent manner, and the various stakeholders, including public health professionals, should be engaged to ensure task shifting is an ethical and effective strategy to improve access to surgical care, as well as promote equity of care. The public, physicians, and the media have to be assured that task shifting will not dilute the quality of care in these settings but will provide much needed surgical services to a population that has been neglected for decades. Neeraja Nagarajan, MD, MPH Varshini Varadaraj, MBBS, MS Author Affiliations: Department of Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland (Nagarajan); Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland (Varadaraj).
jamasurgery.com
(Reprinted) JAMA Surgery July 2015 Volume 150, Number 7
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a Tulane University User on 10/12/2016
687
Our patient population was largely positive for ER, but only 40% received hormonal therapy. Despite this, only 4% of patients experienced a breast cancer–related event. Not performing a sentinel node biopsy may avoid potential morbidity and should be reevaluated for patients 70 years of age or older with clinically negative nodes. Patients were unlikely to have treatment recommendations changed based on a sentinel node biopsy, and adjuvant therapy was less likely to be administered, regardless of nodal status. Patients in this subgroup were more likely to die of causes other than breast cancer, and not performing a sentinel node biopsy did not affect survival. Alice Chung, MD Alexandra Gangi, MD Farin Amersi, MD Xiao Zhang, PhD Armando Giuliano, MD Author Affiliations: Division of Surgical Oncology, Cedars-Sinai Medical Center, Los Angeles, California (Chung, Gangi, Amersi, Giuliano); Department of Surgery, Cedars-Sinai Medical Center, Los Angeles, California (Chung, Gangi, Amersi, Giuliano); Department of Statistics, Samuel Oschin Cancer Center, Cedars-Sinai Medical Center, Los Angeles (Zhang). Corresponding Author: Alice Chung, MD, Division of Surgical Oncology, Department of Surgery, Cedars-Sinai Medical Center, 310 N San Vicente Blvd, 3rd Floor, Los Angeles, CA 90048 ([email protected]). Published Online: May 27, 2015. doi:10.1001/jamasurg.2015.0647. Author Contributions: Drs Chung and Gangi had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Chung, Giuliano. Acquisition, analysis, or interpretation of data: Chung, Gangi, Amersi, Zhang. Drafting of the manuscript: Chung, Gangi, Zhang, Giuliano. Critical revision of the manuscript for important intellectual content: Chung, Amersi, Giuliano. Statistical analysis: Zhang. Obtained funding: Gangi, Giuliano. Administrative, technical, or material support: Chung, Amersi, Giuliano. Conflict of Interest Disclosures: None reported. Funding/Support: The collection, management, and analysis of data for this work was supported by the Margie and Robert E. Petersen Foundation and the Fashion Footwear Charitable Foundation of New York. Role of the Funder/Sponsor: The funders/sponsors had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. 1. Jones EL, Leak A, Muss HB. Adjuvant therapy of breast cancer in women 70 years of age and older: tough decisions, high stakes. Oncology (Williston Park). 2012;26(9):793-801. 2. Rudenstam CM, Zahrieh D, Forbes JF, et al; International Breast Cancer Study Group. Randomized trial comparing axillary clearance versus no axillary clearance in older patients with breast cancer: first results of International Breast Cancer Study Group Trial 10-93. J Clin Oncol. 2006;24(3):337-344. 3. Martelli G, Boracchi P, De Palo M, et al. A randomized trial comparing axillary dissection to no axillary dissection in older patients with T1N0 breast cancer: results after 5 years of follow-up. Ann Surg. 2005;242(1):1-6. 4. Fisher B, Jeong JH, Anderson S, Bryant J, Fisher ER, Wolmark N. Twenty-five-year follow-up of a randomized trial comparing radical mastectomy, total mastectomy, and total mastectomy followed by irradiation. N Engl J Med. 2002;347(8):567-575. 5. Veronesi U, Orecchia R, Zurrida S, et al. Avoiding axillary dissection in breast cancer surgery: a randomized trial to assess the role of axillary radiotherapy. Ann Oncol. 2005;16(3):383-388.
684
6. Hind D, Wyld L, Beverley CB, Reed MW. Surgery versus primary endocrine therapy for operable primary breast cancer in elderly women (70 years plus). Cochrane Database Syst Rev. 2006;(1):CD004272.
COMMENT & RESPONSE
Selective vs Nonselective Nonsteroidal Anti-inflammatory Drugs and Anastomotic Leakage After Colorectal Surgery To the Editor We read with interest the article by Hakkarainen et al1 and wish to commend the authors on their attempt to shed light on the challenging question of nonsteroidal antiinflammatory drugs (NSAIDs) and the risk for anastomotic leakage after colorectal surgery. A growing body of evidence has implicated NSAIDs, particularly selective NSAIDs, as a risk factor for anastomotic leakage. These types of NSAIDs work by inhibiting the cyclooxygenase (COX) class of enzymes: COX1, which is present throughout the body (including the vascular endothelium, stomach, and kidneys) and COX2, which is predominantly found at the site of injury (triggered by inflammatory mediators).2 Hakkarainen et al1 found a 24% increase in the leak rate in their cohort of patients who underwent bariatric surgery, elective colorectal surgery, or nonelective colorectal surgery. In their subgroup analysis, the association was limited to nonelective colorectal surgery (odds ratio [OR], 1.70 [P = .01]), not elective colorectal surgery (OR, 1.13 [P = .36]). They acknowledged that they could not specify the NSAIDs used in their study,1 but they believed that nonselective NSAIDs, such as ketorolac tromethamine, were predominantly used. We feel that there are 2 points requiring emphasis and clarification. First, the classification of selective vs nonselective NSAIDs is often misunderstood. The term selective refers to the newer class of COX2 inhibitors, whereas nonselective NSAIDs are generally thought to indiscriminately inhibit both isoforms. In point of fact, however, this is not the case. Nonselective NSAIDs can inhibit either COX1 or COX2 preferentially. Diclofenac, a nonselective NSAID, behaves like a selective NSAID and, in fact, has an inhibitory profile similar to that of the selective NSAID celocoxib, preferentially inhibiting COX2.2,3 Ketorolac, another nonselective NSAID, preferentially inhibits COX1 more so than any other commercially available NSAID.2 Second, virtually all NSAIDs have the potential to inhibit both COX enzymes regardless of their selectivity. For example, celocoxib will inhibit 60% of COX1 enzymes at 80% of its maximal inhibitory concentration for COX2.2 In our article4 looking at the perioperative use of ketorolac and anastomotic leakage after elective colorectal surgery, we found no association with leakage (OR, 1.21 [P = .66]). We attributed this to ketorolac’s COX1 selectiveness. However, in our subgroup analysis, higher doses were associated with leakage (OR, 1.29 [P = .048]) for every 15-mg increase in ketorolac administered. Perioperative use of NSAIDs and anastomotic leakage remain important issues in the fast-track era of multimodal pain management. Future studies should address the selectivity of
JAMA Surgery July 2015 Volume 150, Number 7 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a Tulane University User on 10/12/2016
jamasurgery.com
Letters
the NSAIDs used and the doses administered over the perioperative period to effectively answer this question. Fady Saleh, MD, MPH Allan Okrainec, MD, MHPE Author Affiliations: Division of General Surgery, University Health Network, Toronto, Ontario, Canada. Corresponding Author: Allan Okrainec, MD, MHPE, Division of General Surgery, University Health Network, 399 Bathurst St, 8-MP 325A, Toronto, ON M5T 2S8, Canada ([email protected]). Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0635. Conflict of Interest Disclosures: None reported. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228. 2. Warner TD, Giuliano F, Vojnovic I, Bukasa A, Mitchell JA, Vane JR. Nonsteroid drug selectivities for cyclo-oxygenase-1 rather than cyclo-oxygenase-2 are associated with human gastrointestinal toxicity: a full in vitro analysis. Proc Natl Acad Sci U S A. 1999;96(13):7563-7568. 3. Klein M, Gögenur I, Rosenberg J. Postoperative use of non-steroidal anti-inflammatory drugs in patients with anastomotic leakage requiring reoperation after colorectal resection: cohort study based on prospective data. BMJ. 2012;345:e6166. 4. Saleh F, Jackson TD, Ambrosini L, et al. Perioperative nonselective non-steroidal anti-inflammatory drugs are not associated with anastomotic leakage after colorectal surgery. J Gastrointest Surg. 2014;18(8):1398-1404.
To the Editor We read with great interest the article by Hakkarainen et al1 published in JAMA Surgery, the results of which are intriguing. The authors report the results of a retrospective cohort study of patients undergoing bariatric or colorectal surgery using data from the Surgical Care and Outcomes Assessment Program linked to the Washington State Comprehensive Abstract Reporting System.1 Although we recognize the importance of this work and applaud the tremendous effort it represents, we would like to draw attention to a number of specific concerns regarding this study.1 An obvious limitation is that the dose, duration, and formulation of the nonsteroidal anti-inflammatory drugs (NSAIDs) are unknown, as are the patient histories of preoperative use of NSAIDs. Although this point is discussed by the authors, we would respectfully question the generalizability of the authors’ assumptions that the “likely” use of intravenous ketorolac tromethamine and caldolor adds support to previous findings showing an increased risk of anastomotic leak with the use of nonselective NSAIDs.2 Another concern is the younger age of the patients receiving NSAIDs because they would be more likely to have a different colorectal pathology than the older patients, and this may also have influenced the results. We also find the rather flexible approach to designs, definitions, outcomes, and analytical modes somewhat worrisome because this type of approach has been implicated in higher rates of “positive findings.”3 We also question why Hakkarainen et al1 felt the need to include heterogeneous groups of patients (ie, patients who underwent bariatric, elective, or nonelective colorectal surgery) in 1 group. We contend that this led to unnecessary subgroup analyses when dealing with colorectal procedures; an indi-
vidual group would have simplified matters. In this context, the wider issue of using large data sets within scientific research and the inherent flaws that can flourish as a result need to be considered.4 Hakkarainen et al1 conclude that the postoperative use of NSAIDs was associated with a significantly increased risk for anastomotic complications among patients undergoing nonelective resection. However, we urge caution and believe that the evidence presented in this study1 does not allow for causation to be established and, therefore, does not convince us to change our current practice of using individualized analgesic regimens. Noel P. Lynch, MCh, MB, Med Sci Emily Boyle, MD Eamon G. Kavanagh, MD, FRCSI Author Affiliations: Department of Surgery, University Hospital Limerick, Limerick, Ireland. Corresponding Author: Noel P. Lynch, MCh, MB, Med Sci, Department of Surgery, University Hospital Limerick, Dooradoyle, Limerick, Ireland (noellynch @rcsi.ie). Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0641. Conflict of Interest Disclosures: None reported. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228. 2. Gorissen KJ, Benning D, Berghmans T, et al. Risk of anastomotic leakage with non-steroidal anti-inflammatory drugs in colorectal surgery. Br J Surg. 2012;99 (5):721-727. 3. Ioannidis JP. Why most published research findings are false. PLoS Med. 2005; 2(8):e124. 4. The PLoS Medicine Editors. Why bigger is not yet better: the problems with huge datasets. PLoS Med. 2005;2(2):e55.
To the Editor I would like to make several comments in relation to the article by Hakkarainen et al1 recently published in JAMA Surgery. Being part of a hospital trust that performs large numbers of colorectal surgical procedures, we have a specific interest in the outcomes of this study.1 The benefits of nonsteroidal anti-inflammatory drugs (NSAIDs) with regard to the opioid-sparing effect cannot be denied. However, some surgeons are concerned about the risk of anastomotic leaks; therefore, they prefer not to use NSAIDs, despite the evidence of leakage being weak and the benefits of NSAID use being very clear. The authors1 acknowledge several limitations, but these do not diminish the use of NSAIDs in any way. When researchers are examining an outcome due to a particular drug, it is of paramount importance that the type of drug (ie, the name of the NSAID), the subtype (cyclooxygenase 1 vs cyclooxygenase 2), the dose, and the duration of therapy are known. Surely, at a clinical level, an anastomotic leak 85 days after surgery cannot be blamed on 2 to 3 doses of NSAIDs given postoperatively? Without any further breakdown of the data on the NSAID therapy that was used, no actual clinical link between the 2 events can be determined. Hakkarainen et al1 actually found a “significant” association between patients with low albumin levels and anastomotic leakage. Few surgeons would argue that patients who
jamasurgery.com
(Reprinted) JAMA Surgery July 2015 Volume 150, Number 7
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a Tulane University User on 10/12/2016
685
Letters
are hypoalbuminemic for whatever reason are at risk of wound breakdown. Are these the same patients who happened to have received NSAIDs? If any conclusion can be drawn from this study,1 it is that, for patients undergoing elective colorectal surgery, NSAIDs appear to be safe and to play an important role in enhanced recovery protocols. We do not agree that there is enough proof to indicate that NSAID use results in anastomotic leakage in patients who have undergone emergency colorectal surgery, but we do agree that NSAIDs should only be used if no other contraindications (such as general debility, renal dysfunction, and dehydration) are indicated. In conclusion, it is clear that proper prospective trials are required to finally answer this question because this study1 poses more questions than it answers. Kevin Doody, BMBS Margaret Coleman, FFARCSI, MSc, SEM Author Affiliations: Department of Anaesthesia, University Hospital Limerick, Limerick, Ireland. Corresponding Author: Kevin Doody, BMBS, Department of Anaesthesia, University Hospital Limerick, Limerick, Ireland ([email protected]). Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0644. Conflict of Interest Disclosures: None reported. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228.
Although a key safety measure, concern over anastomotic leak rates may mask an overall beneficial effect of NSAIDs’ anti-inflammatory and opioid-sparing properties on patient outcomes. Our prospective, multicenter evaluation of NSAID safety in 1500 patients found that administration within 72 hours of surgery was associated with a reduction in overall complications.4 In line with the findings of a previous multicenter study, we found no evidence of an association between mixed NSAIDs (predominantly ibuprofen) and anastomotic leak.5 The absolute risk difference in anastomotic leak between NSAID and non-NSAID groups in Hakkarainen and colleagues’ study was 0.6%. The clinical significance of this small effect size set against the potential benefits of NSAIDs remains to be determined. Three multicenter observational studies have demonstrated the safety of NSAIDs and, particularly, ibuprofen in elective colorectal patients. The existing evidence base is unlikely to be enhanced by further observational studies. A large randomized clinical trial is required to test the efficacy and safety of NSAIDs in reducing complication rates following elective colorectal surgery. Thomas M. Drake, BMedSci Dmitri Nepogodiev, MBChB Henry A. Claireaux, BSc(Hons) Author Affiliations: Student Audit and Research in Surgery (STARSurg) Steering Group, Birmingham, England. Corresponding Author: Dmitri Nepogodiev, MBChB, Academic Department of Surgery, Old Queen Elizabeth Hospital, Room 29, 4th Floor, Edgbaston, Birmingham B15 2TH, England ([email protected]).
To the Editor We commend Hakkarainen and colleagues1 for their contribution to the ongoing discussion regarding the association of nonsteroidal anti-inflammatory drugs (NSAIDs) and anastomotic leak following gastrointestinal surgery. A previous meta-analysis found NSAIDs are significantly associated with an increased risk for anastomotic leak.2 To our knowledge, the study by Hakkarainen et al is the first to suggest that this effect is limited to emergency colorectal patients only. These findings are based on a well-conducted analysis of clearly defined groups from a large prospective database. However, the patients in the NSAID group were significantly younger, with fewer comorbidities than the control arm, which makes it difficult to generate reliable effect estimates between treatment groups. Comparison between groups is further hindered by a lack of data on perioperative and postoperative variables known to be associated with anastomotic leak.3 Nonetheless, it would be interesting to perform a casematched analysis of the data from Hakkarainen et al and use factors associated with anastomotic leakage to generate improved effect estimates. The potency and adverse effect profiles of individual NSAID agents vary greatly. Data regarding choice of NSAID agent was unavailable in the study by Hakkarainen and colleagues. If correct, their assumption that most patients received ketorolac would limit the generalizability of this study to centers outside of the United States, many of which use ibuprofen in preference to ketorolac. 686
Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0806. Conflict of Interest Disclosures: None reported. Additional Information: The STARSurg Collaborative Steering Group members include Thomas M. Drake, Dmitri Nepogodiev, Henry Claireaux, Midhun Mohan, Chetan Khatri, Chia Kong, Lisa McNamee, Michael Bath, Stephen J. Chapman, James Glasbey, Ewen M. Harrison, Aneel Bhangu, and J. Edward Fitzgerald. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228. 2. Bhangu A, Singh P, Fitzgerald JE, Slesser A, Tekkis P. Postoperative nonsteroidal anti-inflammatory drugs and risk of anastomotic leak: meta-analysis of clinical and experimental studies. World J Surg. 2014;38(9): 2247-2257. 3. McDermott FD, Heeney A, Kelly ME, Steele RJ, Carlson GL, Winter DC. Systematic review of preoperative, intraoperative and postoperative risk factors for colorectal anastomotic leaks. Br J Surg. 2015;102(5):462-479. 4. STARSurg Collaborative. Impact of postoperative non-steroidal anti-inflammatory drugs on adverse events after gastrointestinal surgery. Br J Surg. 2014;101(11):1413-1423. 5. Klein M, Gögenur I, Rosenberg J. Postoperative use of non-steroidal anti-inflammatory drugs in patients with anastomotic leakage requiring reoperation after colorectal resection: cohort study based on prospective data. BMJ. 2012;345:e6166.
In Reply We are pleased that our study1 has generated debate and discussion regarding the perioperative use of nonsteroidal anti-inflammatory drugs (NSAIDs) among patients undergoing gastrointestinal surgery, and we sincerely thank the several teams of physicians and surgeons who have taken the time to respond. First, we would like to reiterate that we do not sug-
JAMA Surgery July 2015 Volume 150, Number 7 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a Tulane University User on 10/12/2016
jamasurgery.com
Letters
gest that our data firmly establish causality between perioperative use of NSAIDs and increased rates of anastomotic leak. We believe that our data are strong enough to warrant increased caution when deciding whether or not to use NSAIDs for patients undergoing emergency colorectal procedures. Nor do we dispute the possible benefits of NSAID use in enhanced recovery protocols for their opioid-sparing potential. It is important to note that our data and conclusions are focused on patients undergoing nonelective colorectal procedures, and thus the authors who defend the use of NSAIDs in recovery protocols may be reading too much into our conclusions and taking them out of context. We agree with all the authors who responded that further studies are warranted and needed to address this question and to analyze the type of drug used, the dosage, and the duration of therapy. In our study,1 we simply urge caution. Regarding the selectivity of various NSAIDs, we agree with those authors who commented that all NSAIDs are selective for either cyclooxygenase 1 (COX1) or COX2 to varying degrees at different doses. Both ketorolac tromethamine and ibuprofen are relatively nonselective at the doses typically given, with only a mildly increased inhibition of COX1 vs COX2, and we consider them nonselective. Despite the lack of specific drug information in the Surgical Care and Outcomes Assessment Program (SCOAP) registry, we are quite confident that ketorolac and/or ibuprofen represent the large majority of NSAIDs used. We can be this confident because of the collaborative nature of the SCOAP sites and because of our ability to directly query surgeons at participating sites about formularies and practice patterns. As to our initial inclusion of patients undergoing bariatric or elective colorectal procedures, this is a function of our initial interest in the potential effects of NSAIDs on gastrointestinal anastomotic healing. Not including data from these 2 patient populations in our study 1 could be construed as disingenuous because it would be different from the original research question that we set out to answer, and, obviously, the data from these 2 patient populations yield different results from analysis of the patients undergoing nonelective colorectal surgery. Finally, an error regarding the sample size referenced in the title of our Table 3 occurred during the proofing process. The sample size of the cohort of patients undergoing nonelective colorectal procedures was 1621 patients, not 9624. This error will be corrected online. Timo W. Hakkarainen, MD, MS David R. Flum, MD, MPH Author Affiliations: Department of Surgery, University of Washington Medical Center, Seattle. Corresponding Author: Timo W. Hakkarainen, MD, MS, Department of Surgery, University of Washington Medical Center, 1959 NE Pacific St, PO Box 356410, Seattle, WA 98105-6410 ([email protected]). Published Online: May 6, 2015. doi:10.1001/jamasurg.2015.0638. Conflict of Interest Disclosures: None reported. 1. Hakkarainen TW, Steele SR, Bastaworous A, et al. Nonsteroidal anti-inflammatory drugs and the risk for anastomotic failure: a report from Washington State’s Surgical Care and Outcomes Assessment Program (SCOAP). JAMA Surg. 2015;150(3):223-228.
Helping Meet Surgical Needs in Under-resourced Settings: The Role of Task Shifting To the Editor We commend the timely article by Kotagal and Horvath1 that discusses the challenges of providing surgical care in under-resourced settings, and we look forward to more articles that address this important issue. As the authors point out, low- and middle-income countries have complex health systems with major discrepancies in quality of and access to care even within each country/region. In many low- and middle-income countries, private medical care contributes the lion’s share (80%) of the services provided, putting it out of the reach of millions.2 In a middleincome country like India, the overall physician to population ratio is 1:1800 (below 1:1000 established by the World Health Organization), while the situation in the rural areas of India is dire with only 25% of trained physicians available for 60% to 70% of the population. The Indian government reports that 70% of specialist (surgeons, pediatricians, and gynecologists) positions in rural government health centers are vacant. Proposals by governments to make rural service a mandatory component of medical education have been met with resistance by physicians and their associations. Surgeons and other specialists are reluctant to work in rural areas, pointing out a lack of infrastructure and opportunities for professional growth, among other reasons. As a way to break this stalemate, low- and middleincome countries have started programs that are geared toward training midlevel health care professionals to serve in rural and underserved areas. While these programs have traditionally been focused on managing medical conditions, training these health care professionals in basic obstetric and surgical care would be hugely advantageous in reducing preventable mortality and morbidity. The training should include recognition of signs and symptoms of common conditions needing surgical consultation, as well as management of simple surgical and obstetric emergencies. Task shifting and creating a cadre of skilled midlevel health care professionals for basic surgical care have been validated in studies3-5 from under-resourced settings that have shown that the outcomes associated with these programs are comparable to the outcomes associated with the care provided by physicians. The potential limitations of task shifting need to be addressed in a diligent manner, and the various stakeholders, including public health professionals, should be engaged to ensure task shifting is an ethical and effective strategy to improve access to surgical care, as well as promote equity of care. The public, physicians, and the media have to be assured that task shifting will not dilute the quality of care in these settings but will provide much needed surgical services to a population that has been neglected for decades. Neeraja Nagarajan, MD, MPH Varshini Varadaraj, MBBS, MS Author Affiliations: Department of Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland (Nagarajan); Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland (Varadaraj).
jamasurgery.com
(Reprinted) JAMA Surgery July 2015 Volume 150, Number 7
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archsurg.jamanetwork.com/ by a Tulane University User on 10/12/2016
687
