A 75-year-old man with a past medical history of end-stage renal disease (on hemodialysis), hypertension, and diabetes was found to have obstructive, multi-vessel coronary artery disease. The patient was referred for bypass surgery and mitral valve repair. He underwent successful surgery and was discharged to a rehabilitation facility where he recuperated at the expected rate.

The patient continued to attend hemodialysis and clinic visits without fail. Eight months later, he was admitted to the hospital with shaking chills and confusion and found to have an empyema (i.e., collection of pus) on the left side of his chest. He underwent thoracic surgery where the trapped lung was freed from its adhesions and a 12-cm red-rubber snare was recovered—a surgical instrument from the prior surgery. The patient recovered to his baseline level of functioning after a 10-day intensive care unit (ICU) course followed by treatment in a rehabilitation facility.

On subsequent investigation, hospital staff members were puzzled by the failure to detect the object. The patient underwent three-times weekly dialysis sessions, at least eight clinic visits with multiple physicians, and had no symptoms until his dramatic presentation many months later. He also had seven chest radiographs postoperatively in which the device was not detected.

by Robert R. Cima, MD, MA

Retained surgical items (RSIs), formerly described as retained foreign objects, are an extreme and obvious example of a medical error (Figure). In legal parlance, these cases are described by the doctrine of "res ipsa loquitur" which is Latin for "the thing speaks for itself."(1) This means that a breach in duty or care can be inferred directly from the very nature of the event without knowing the circumstances, intent, or actions of the defendant. Unintentionally leaving a surgical instrument or item within a patient is simply not defensible. This principle is the basis of the National Quality Forum inclusion of RSIs as one of the original 10 "Never Events."

The exact incidence of RSIs is unknown. Values for these events range from approximately every 5500 to 18,000 surgical procedures.(2-7) However, some estimates are as frequent as 1 in every 1000 abdominal operations.(7) The number of patients potentially affected by an RSI is sizable based on the 2007 estimated annual volume of both outpatient and inpatient surgeries in the United States of slightly greater than 20 million procedures.(8) Because many RSI cases are kept confidential by providers, institutions, and insurance companies, the exact number of patients affected is unknown, but a conservative estimate would be around 1000 cases a year in the US. RSIs can result in significant patient morbidity and even mortality.(3,9,10) Patients can experience immediate postoperative complications such as infections or bowel obstruction. However, many RSIs present in a delayed fashion. In the case above, an RSI went undetected for 8 months. There are many case reports of patients who are asymptomatic or live with chronic unexplained symptoms for years or even decades after the index operation before the RSI is discovered.(11,12)

Gawande's study of RSIs identified three risk factors: emergency operations, unplanned changes in the operation, and patients with higher body mass index (BMI).(3) However, in the largest single institution study of more than 200,000 procedures, the 34 recorded RSIs were not associated with any of these risk factors.(4) In fact, all of the RSI events occurred in routine uncomplicated cases. Operations involving the chest or abdomen are more commonly associated with RSIs, but RSIs can occur after almost any procedure and involve almost all operative sites.(6) Subsequent studies have identified additional risk factors for RSIs (Table).(13)

Although there are published operating room guidelines and recommended procedures to account for surgical instruments and sponge products used during surgery, their effectiveness in preventing RSIs is limited.(14) Unfortunately, count discrepancies are fairly common. In an observational study, Greenberg and colleagues demonstrated that a count discrepancy occurred in 1 of every 8 operations.(15) This study identified count discrepancies that were recognized at the time of the operation resulting in activation of procedures to resolve the discrepancy. However, a consistent finding in nearly all the published reports of RSIs was that the instrument and sponge counts were reported as "correct" prior to closure in the majority of cases.(3,4,16) These sobering data illustrate that counting surgical instruments and sponges is more complex than it might appear.

All the case series describing RSIs indicate that cotton sponges are the most common RSIs.(3,4) Many factors contribute to the difficulty accounting for sponges (as opposed to instruments). Unlike the majority of instruments, sponges are used in very high numbers during a case. They are used by many members of the operating team and, unlike instruments, frequently leave the hands of the team members for (temporary) insertion into the patient's body. During the course of an operation, sponges can move quite easily from one place to another inside a body cavity as the result of patient position change, manipulation of the body cavity contents, or movement of organs, especially the intestines. Additionally, over the course of a surgery, the sponge color and size changes dramatically making it more difficult to identify within a cavity. Lastly, counting sponges removed from the operating site over the course of a procedure is associated with "simple counting" errors and is further compounded by different personnel performing the counts, competing activities, distractions, and poor intra-team communication.

To address this problem, technology-based sponge counting and identification systems have been employed with striking success. Some options are as simple as hanging clear pouches to collect sponges to labeling sponges with data matrix tags (DMT) or even radiofrequency identification chips (RFID). The more advanced technologies, DMT or RFID, are a counting or a counting and detecting system, respectively. Using DMT technology is similar to the barcode reader used at the local grocery store. Each individual sponge has a unique barcode label. A reader is used to scan the sponge onto the operative field and then the sponge is rescanned when it is removed from the field. The reader then acts as a ledger so that at the end of the case all sponges scanned in have to be scanned out of the case. In the RFID systems, a radiofrequency tag is embedded in each sponge. They are scanned in and out using the RFID as opposed to a barcode. In some RFID technology sponges, a wand can be used to detect a sponge if it has not been accounted for. In the largest reported experience using the DMT system, the Mayo Clinic reported that introduction of data matrix–labeled sponges eliminated cotton sponge RSIs over an 18-month period, while previously they were experiencing one such case every 76 days.(17) A small study showed that the RFID system could detect a retained sponge during an operation with 100% sensitivity and 100% specificity.(18) Recently, a prospective study of 2285 patients also demonstrated the clinical functionality of RFID-labeled sponges.(19)

These advanced technologies add to the cost of surgical care to a varying degree. However, the "cost" of an RSI is not only borne by the affected patient. These events can have significant financial impact on providers and institutions, both in legal costs and nonpayment from the federal government.(16,20) Additionally, institutions suffer a cost to their reputation because in many states RSIs are mandated to be reported as part of public reporting initiatives. There are as yet no studies that have weighed these relative costs and determined the return on investment for advanced technology systems to prevent RSIs. There are no studies that have evaluated the patient safety and economics risk–benefit of introducing this new technology. It is unclear given the very low frequency of RSIs and rare public disclosure of such events that sufficient data could be accumulated to make a broad public health recommendation. However, analysis of institutional costs over time might provide some empirical evidence on which to base a decision.

RSIs are fortunately an uncommon but profoundly disturbing surgical error. RSIs expose the patient to increased morbidity and mortality. Technological solutions are available to address the most common type of RSI, cotton sponges, but they are not widely used due to ongoing cost, the low frequency of events, and competing institutional priorities. It is unlikely that RSIs will be eliminated easily in the near future. At present, the best defense against RSIs is following standardized counting policies, attention to detail by all operating room team members, improved team communication, and, where possible, application of technology. Extensive education and team-based approaches have been shown to reduce but not eliminate sponge or instrument RSIs.(21) However, key findings from a number of studies on operating room safety suggest improving team communication, flattening hierarchy, and teaching and enhancing teamwork lead to fewer errors in the operating room. The appropriate addition of technology might augment the human factor elements of operating room performance.

Take-Home Points

• An RSI is an uncommon but profoundly disturbing surgical error.
• RSIs expose the patient to increased morbidity and mortality.
• The traditional solution to preventing RSIs has been standardized counting policies blending with heightened attention to detail by all operating room team members and improved communication. However, such strategies have only limited effectiveness.
• Recently, new technologies such as DMT- and RFID-labeled sponges have demonstrated high levels of effectiveness in preventing RSIs. Such technologies are costly, and further clinical and economic analyses should be done to determine the return on investment for individual hospitals and surgical centers.

Robert R. Cima, MD, MA
Vice-chair, Quality and Safety, Department of Surgery
Associate Professor of Surgery, Mayo Clinic College of Medicine
Consultant, Division of Colon and Rectal Surgery

1. Legal Dictionary. Law.com. [Available at]

2. Gonzalez-Ojeda A, Rodriguez-Alcantar DA, Arenas-Marquez H, et. al. Retained foreign bodies following intra-abdominal surgery. Hepatogastroenterology. 1999;46:808-812. [go to PubMed]

3. Gawande AA, Studdert DM, Orav EJ, Brennan TA, Zinner MJ. Risk factors for retained instruments and sponges after surgery. N Engl J Med. 2003;348:229-235. [go to PubMed]

4. Cima RR, Kollengode A, Garnatz J, Storsveen A, Weisbrod C, Deschamps C. Incidence and characteristics of potential and actual retained foreign object events in surgical patients. J Am Coll Surg. 2008;207:80-87. [go to PubMed]

5. Rappaport W, Haynes K. The retained sponge following intra-abdominal surgery. A continuing problem. Arch Surg. 1990;125:405-407. [go to PubMed]

6. Lincourt AE, Harrell A, Cristiano J, Sechrist C, Kercher K, Henifort BT. Retained foreign bodies after surgery. J Surg Res. 2007;138:170-174. [go to PubMed]

7. Hyslop JW, Maull KI. Natural history of the retained surgical sponge. South Med J. 1982;75:657-660. [go to PubMed]

8. Russo CA, Elixhauser A, Steiner C, Wier L. Hospital-Based Ambulatory Surgery, 2007. HCUP Statistical Brief #86. Rockville, MD: Agency for Healthcare Research and Quality; February 2010. [Available at]

9. Yildirim S, Tarim A, Nursal TZ, et al. Retained surgical sponge (gossypiboma) after intraabdominal or retroperitoneal surgery: 14 cases treated at a single center. Langenbecks Arch Surg. 2006;391:390-395. [go to PubMed]

10. Falleti J, Somma A, Baldassarre F, Accurso A, D'Ettorre A, Insabato L. Unexpected autoptic finding in a sudden death: gossypiboma. Forensic Sci Int. 2010;199:e23-e26. [go to PubMed]

11. Taylor FH, Zollinger RW 2nd, Edgerton TA, Harr CD, Shenoy VB. Intrapulmonary foreign body: sponge retained for 43 years. J Thorac Imaging. 1994;9:56-59. [go to PubMed]

12. Szentmariay IF, Laszik A, Sotonyi P. Sudden suffocation by surgical sponge retained after a 23-year old thoracic surgery. Am J Forensic Med Pathol. 2004;25:324-326. [go to PubMed]

13. Wan W, Le T, Riskin L, Macario A. Improving safety in the operating room: systematic literature review of retained surgical sponges. Curr Opin Anaesthesiol. 2009;22:207-214. [go to PubMed]

14. Goldberg JL, Feldman DL. Implementing AORN recommended practices for prevention of retained surgical items. AORN J. 2012;95:205-216. [go to PubMed]

15. Greenberg CC, Regenbogen SE, Lipsitz SR, Diaz-Flores R, Gawande AA. The frequency and significance of discrepancies in the surgical count. Ann Surg. 2008;248:337-341. [go to PubMed]

16. Kaiser CW, Friedman S, Spurling KP, Slowick T, Kaiser HA. The retained surgical sponge. Ann Surg. 1996;224:79-84. [go to PubMed]

17. Cima RR, Kollengode A, Clark J, et al. Using a data-matrix-coded sponge counting system across a surgical practice: impact after 18 months. Jt Comm J Qual Patient Saf. 2011;37:51-58. [go to PubMed]

18. Macario A, Morris D, Morris S. Initial clinical evaluation of a handheld device for detecting retained surgical gauze sponges using radiofrequency identification technology. Arch Surg. 2006;141:659-662. [go to PubMed]

19. Rupp CC, Kagarise MJ, Nelson SM, et al. Effectiveness of a radiofrequency detection system as an adjunct to manual counting protocols for tracking surgical sponges: a prospective trial of 2,285 patients. J Am Coll Surg. 2012 Jul 5; [Epub ahead of print]. [go to PubMed]

20. Baltimore, MD: Department of Health & Human Services, Centers for Medicare & Medicaid Services; July 2008. SMDL #08-004. [Available at]

21. Cima RR, Kollengode A, Storsveen AS, et al. A multidisciplinary team approach to retained foreign objects. Jt Comm J Qual Patient Saf. 2009;35:123-132. [go to PubMed]

Table. Selected risk factors for retained surgical items.(13)

Figure. A retained malleable retractor after an abdominal exploration and tumor debulking. It was detected in the immediate postoperative period after a chest radiograph for intraoperative line placement was performed.