Natural historyneed for team tretrieval

.DR.N.N.ni,D.g

Susan D. Moffatt-Bruce, M.D., Ph.D. , OPUS 12 Foundation Multicenter Trials Group

University School of Medicine, Philadelphia, PA, USA

(n5 13), a total of 71 occurrences were analyzed for (1) item location and type; (2) time to presentation)

l

The authors declare no external financial or non-financial support with regards to this work. None of the authors received grants, honoraria, consultancies,

speakers bureau or advisory-board positions, or significant stock holdings in connection to or as a result of this work.

.

The American Journal of Surgery (2014) -, --* Corresponding author. Tel.: 11-614-293-1964; fax: 11-614-293-9155and/or discovery; (3) presenting signs and symptoms; (4) procedure and incision characteristics; (5pathology reports; and (6) patterns of SOVs abstracted from medical and operative records. TheseSOV were then grouped into individual vs team errors and single- vs multifactorial occurrences.

RESULTS: Among 71 cases, there were 48 women and 23 men. Mean patient age was 49.7 617.5 years (range 19 to 83 years). Mortality was 4 of 71 (5.63%, only 1 attributable to RSI). Twelvecases (16.9%) occurred at nonparticipating referring hospitals. Most RSI procedures (62%) occurredon the day of hospital admission. The median time from index RSI case to retained item removaKEYWORDS:Retained surgicalitems;Natural history;Intraoperativecausative factors;Team patient safety

AbstractBACKGROUND: Unintentionally retained items feature prominently among surgical never events.

Our knowledge of these rare occurrences, including natural history and intraoperative safety omission orvariance (SOV) profile, is limited. We sought to bridge existing knowledge gaps by presenting a second-ary analysis of a multicenter study focused on these important aspects of retained surgical items (RSIs).

METHODS: This is a post hoc analysis of results from a multicenter retrospective study of RSIsbetween January 2003 and December 2009. After excluding previously reported intravascular RSIsMedicine, New Brunswick, NJ, USA; eDepartment of SurgeraDepartment of Surgery, The Ohio State University College of Medicine, Columbus, OH, USA; bDepartment of Surgery,Miami Valley Hospital/Wright State University School of Medicine, Dayton, OH, USA; cDepartment of Surgery, St LukesUniversity Health Network, Bethlehem, PA, USA; dDepartment of Surgery, UMDNJ-Robert Wood Johnson School of

y, Vanderbilt University Medical Center, Nashville, TN, USA;fDepartment of Surgery, Cooper University Hospital, Camden, NJ, USA; gDepartment of Surgery, Thomas JeffersonS. Peter Stawicki, MLaurie Chowayou,Susette M. Coyle, RRaffaele MarchigiaNiels D. Martin, M.E-mail address: stawicki.ace@gmail

Manuscript received July 21, 2013;

0002-9610/$ - see front matter 2014http://dx.doi.org/10.1016/j.amjsurg.20of retained surgical items supports theraining, early recognition, and prompt

.a,*, Charles H. Cook, M.D.a, Harry L. Anderson, III, M.D.b,., B.S.N.b, James Cipolla, M.D.c, Hesham M. Ahmed, M.D.d,, M.S.N.d, Vicente H. Gracias, M.D.d, David C. Evans, M.D.a,M.D.c, Raeanna C. Adams, M.D.e, Mark J. Seamon, M.D.f,, Steven M. Steinberg, M.D.a,

a.com

revised manuscript September 12, 2013

Elsevier Inc. All rights reserved.

13.09.029

0 daicatioon ction5 22ts, m). IsothatVs ees phe i

s rese

this study overlap with data pinitial study of RSI risk factorsfocus of the aforementioned mthe majority of the data incluAppropriate annotations are pwhenever an overlap exists withdata. For the purposes of descriindex RSI procedure and the iditem, the following definitions(within 24 hours), acute (.2acute (between 1 week and

issi

2 The American Journal of Surgery, Vol -, No -, - 2014(.6 weeks).Intraoperative safety omwere abstracted from patient recreviously reported in our,8 although the scope andanuscript8 did not includeded in the current report.laced throughout the textany previously published

bing the time between theentification of the retainedwere used: immediate4 hours to 1 week), sub-6 weeks), and chronic

ons or variances (SOVs)

present study provides unique insight into factors and/orevents that may have led to RSI occurrences in this series.

For SOV category involving RSI missed on imaging,all available information regarding the particular eventwere reviewed by each reporting sites principal investi-gator. Specific information sought during this processincluded (1) availability of any communication regardingpreliminary vs final radiology interpretation; (2) anymention of radiographs being examined by the surgeon/surgical team; (3) any mention within the medical record(ie, operative nursing or surgical report) of radiographyresults being communicated to the operating room team;and (4) any available evidence from corresponding surgicalthe nature of the current report, some of the data used in the authors postulate that despite this shortcoming, theThis report represents a post hoc examination of datafrom a 7-center retrospective study of RSI. Original data for84 RSI cases were abstracted from January 2003 andDecember 2009. After excluding previously reportedintravascular RSIs6 (n 5 13), a total of 71 RSI occurrenceswere analyzed for (1) item location/type; (2) time to pre-sentation/discovery; (3) presenting signs/symptoms; (4)procedure/incision characteristics; and (5) surgical pathol-ogy reports. These 71 occurrences include both RSI thatoccurred in participating centers and RSI that occurred atnonstudy hospitals that were surgically removed at partici-pating study hospitals during the study period. Inherent towas 2 days (range ,1 to .3,60was the most common identifasymptomatic. The most comm(n 5 22), abscess/fluid collecincluded exudative reaction (nreview of intraprocedural evenand 2 or more SOVs (37 of 71

CONCLUSIONS: The findingsystem errors and 2 or more SOthat early RSI removal minimizidentification and treatment. Twith longer retention periods. 2014 Elsevier Inc. All right

Unintentional retained items feature prominently amongthe surgical never events or adverse clinical occurrencesthat are broadly considered unacceptable and felt to betotally preventable.1 Details regarding natural history andintraoperative events related to retained surgical items(RSIs) remain limited and only consist of clinical reviews,case series, and isolated reports.17 This study aims toanswer many of the outstanding questions regarding clin-ical signs, symptoms, diagnostic evaluation, anatomic loca-tions, and intraoperative characteristics associated withRSIs.

Methodsords (Table 1 and Fig. 1).ys, n 5 63). Abdominal RSIs predominated, and plain radiographyn method. Most RSIs removed early (,24 hours, n 5 23) werelinical/diagnostic findings in the remaining group were focal pain(n 5 18), and mass (n 5 8). Most common pathology findings), fibrosis (n 5 17), and purulence/abscess (n 5 15). On detailedost RSI cases were found to involve team/system errors (50 of 71)lated human error was seen in less than 10% of cases.most operations complicated by RSIs were found to involve team/mphasizes the importance of team safety training. The observationatient morbidity and symptoms highlights the need for prompt RSIncidence of inflammation-related findings increases significantly

rved.

In addition to reviewing their own institutions operative/perioperative records, each site reviewed previous operativereports for cases originating at other institutions. Occur-rences were categorized as either individual or team/systemerrors and further grouped by the total known number ofknown SOVs per case. Isolated human error was definedas an error clearly attributable to single individuals actions.Team or system error was defined as a combination of (1)error not isolated to single individual actions; (2) errorattributable to 2 or more co-associated SOVs; (3) errorinvolving insufficient safety cross-checks/redundancy; (4)error involving lack of safety knowledge/education; and(5) errors in safety verification, documentation, or commu-nication (if/when known).

Regarding procurement of specific event-related infor-mation, each occurrence was reviewed by the respectivesite investigator at every contributing institution beforebeing transmitted to the central location. This processinvolved one or more of the following: (1) careful review ofthe medical record; (2) surgical quality or sentinel eventquery review, if available; (3) any other internal reportsavailable on record (ie, anonymous complaints, etc); and(4) the presence vs absence and the observance of pertinentsafety protocols, techniques, and/or other measures. Giventhe very nature of what information was available, thesubjectivity involved, and the mode of abstraction, there islikely a significant amount of under-reporting. However,quality/sentinel event queries, if available. Similarly, for the

category RSI missed on tagging, all available documents(including medical records, surgical quality, and sentinelevent queries) were reviewed for (1) the use of taggingequipment at the time of event; (2) information regarding

(Table 2). Clinical signs/symptoms accessible to study in-vestigators were reviewed by each participating sites prin-cipal investigator. For patients with more than 1 clinicalsign/symptom, each additional finding was listed under cor-responding, separate category. Similar procedure was fol-lowed for pathology findings, where cases with more than1 abnormal result having each finding were listed underseparate category (Table 3).

Descriptive statistics were employed to tabulate and reportthe data. Results are presented as proportions and centraltendencies of the data and are compiled into graphs andtables. Differences in categorical variables were analyzedusing chi-square or Fisher exact test, as appropriate. Statis-tical significance was considered at alpha less than .05.Minitab 16 (Minitab, Inc., State College, PA) software wasused to carry out statistical procedures.

Results

Descriptive characteristics of the retainedsurgical item sample

Table 1 Summary of SOV as determined during the review ofrecorded medical/operative records for the 71 study patients.Note that some cases involve more .1 SOV and that somecases involve .1 SOV in a single category

Safety variance/omission Patients %

Lack of safety steps verificationin medical record

28 39.4

RSI missed on initial imaging/tagging interpretation

20 28.2

Count not performed or documented 12 16.9Protocol followed incorrectly(ie, out of sequence)

11 15.5

Protocol disregarded 9 12.7Lack of protocol/safety procedure knowledge 5 7.04Team communication inadequate/deficient 4 5.63Inadequate device/item tracking 3 4.23

RSI 5 retained surgical item; SOV 5 safety omissions or variances.

S.P. Stawicki et al. RSI: Need for team training, early recognition, and prompt retrieval 3staff education/training (surgical quality/sentinel eventqueries, if available); (3) evidence of established scanningprocedures being followed or not followed; and (4) evi-dence of scanning equipment failure. Again, despite likelyunder-reporting bias, the authors believe that any informa-tion gathered in this fashion provides valuable glimpsesinto the anatomy or RSI events.

To obtain information related to presenting signs/symptoms, medical records were reviewed in detail andclinical signs/symptoms were grouped into broad categoriesFigure 1 Coincidence of events and RSI occurrence. Index events are lrow/column represents co-occurrence of RSI-related intraoperative SOVRSI were missed by imaging/tagging, 5 also had protocols disregardedmissed on initial imaging/tagging. Co-associated events are color codoutnumber the index event for each respective row/column, then multiplegory. Conversely, if the total number of co-associated events is fewer thathen more index occurrences than co-occurrences were noted for that cdirectly or indirectly related to team factors and/or lack of team cross-Among 71 total cases, there were 48 women and 23men. The mean patient age was 49.7 6 17.5 years (range19 to 83 years). Mortality was 4 of 71 (5.63%). Onemortality (1.41%) was deemed directly attributable to RSIand was associated with postoperative abdominal sepsisafter RSI retrieval procedure (exploratory laparotomy).Twelve cases (16.9%) originated at nonparticipating refer-ring hospitals. Most RSIs were identified and removedduring the index hospital admission (42 of 71), thusrequiring readmission for RSI in 29 of 71 (40.8%) cases,

abeled in black and are considered denominators. Each intersectings and are, therefore, numerators. For example, of 20 cases where. Likewise, for 9 patients with protocols disregarded, 5 had RSIed according to increasing event count. If co-associated eventsco-associated events were present for the index event in that cate-n the number of index occurrences for the respective row/column,ategory. * Denotes SOV categories that were frequently either

verification.

Most procedures complicated by RSI (44 of 71 or 62%)

Little to no reaction 21 29.6were day of admission surgeries. The remaining RSIprocedures (38%) were performed on subsequent postad-mission days. Most RSI procedures occurred early duringhospitalization (median 1 day, mean 4.1 days). The medianwith 2 additional patients (2.82%) requiring second read-mission related to RSI. The median length of stay forreadmissions was 3 days (range 1 to 80 days).

Timing of retained surgical item procedures andretained surgical item identification/retrieval

syndromeIleus 7 10.6Respiratory 6 9.09Peripheral pain 5 7.58Lack of expected clinical progress 5 7.58Fluid collection 4 6.06None 23 34.8

A significant proportion of patients reported more than 1 symptom,

and some overlap between categories is present.

*Data available for 66 patients.Table 2 Overall distribution of symptoms in the studysample

SymptomNumber ofpatients*

%Cases*

Abdominal pain 17 25.8Abscess 14 21.2Nausea/vomiting 10 15.2Wound complication 10 15.2Mass 8 12.1Systemic inflammatory response 7 10.6

4time from index RSI occurrence to RSI removal was 2 days(range ,1 day to 10 years, n 5 63). Twenty-three patientsunderwent immediate (ie, within 24 hours) RSI removal,and more than 80% of RSI were retrieved within 3 monthsof the index operation (Fig. 2). Patients with abdominal/pelvic locations of RSI had significantly longer intervalsto retained item identification/removal (median 4 days,range 0 to 885 days) than patients with extra-abdominal/pelvic RSI (median 1 day, range 0 to .3,600 days,MannWhitney U test, P , .02).

Clinical signs and symptoms

The overall distribution of signs/symptoms associatedwith RSI is shown in Table 2. The proportion of patients withclinically significant symptoms was not statistically differentbetween the textile (ie, surgical sponge) and nontextile(ie, surgical instrument) retained items (64% vs 66%, respec-tively, P . .05). The duration of time between the index RSIcase and the recognition/removal of the retained item, how-ever, correlated with clinical symptoms. Most RSIs removedwithin the first 24 hours were asymptomatic (20 of 23 or87%). Longer RSI retention was associated with progres-sively greater proportion of symptomatic patients, with thoseclassified as chronic having symptoms in 16 of 19 or 84%cases, those classified as subacute in 8 of 10 or 80% cases,those listed as acute in 11 of 14 or 79% cases, and thosein the immediate group in 3 of 23 or 13% cases (data avail-able for 66 of 71 cases, chi square, P , .01).

Roughly half of patients had co-occurrence of more than1 symptom (32 of 66 or 48.5%). RSIs located in theabdomen/pelvis were more likely to have clinically signif-icant presenting symptoms (29 of 43 or 67%) than non-abdominal/pelvic RSI (9 of 23 or 39%, Fisher exact test, P,.04). Among 43 patients with abdominal/pelvic RSIs, mosthad abdominal pain (n 5 17), abscess/mass (n 5 16),nausea/vomiting (n 5 9), and ileus (n 5 7). More nonspe-cific complaints/findings included systemic inflammatory

Retained object (grossdescription only)

22 31.0

A number of reports demonstrated more than 1 pathologic

descriptor, and some overlap between categories is present.Table 3 Overall distribution of pathology findings in thestudy sample

Pathology description/finding Number of patients % Cases

Fibrotic reaction 23 32.4Exudative reaction 22 31.0Purulence/abscess 13 18.3Acute inflammation 11 15.5Mass (gross description only) 6 8.5Granuloma 4 5.6Tissue erosion by the retainedobject

3 4.2

Chronic inflammation 3 4.2Fistula tract/sinus 3 4.2

The American Journal of Surgery, Vol -, No -, - 2014response syndrome (n 5 5) and lack of expected clinicalprogress (n 5 5). For combined extremity, head/neck, nat-ural orifice, and thoracic cases (n5 23), localized pain at theRSI site predominated (n5 5), followed by abscess/mass (n5 3) and systemic inflammatory response syndrome (n5 2).

Diagnostic imaging and retained surgical itemlocation

Plain imaging and computed tomography (CT) werethe most common localizing modalities. Most RSIs werediagnosed by plain films of the operative anatomic area (36or 51% cases). Because of limited data in medical record, itwas difficult to clearly categorize such films into non-incidental (ie, symptom or suspicion driven) vs incidental(unrelated to surgical procedure). Nonincidental CTs or-dered to confirm the suspected RSIs on plain films wereperformed in 24 cases (34%), whereas purely incidentalRSI findings on CT were seen in only 2 cases (2.8%, bothchronic RSI retention cases). Other radiographic diagnostic

Figure 2 Temporal representation of the raw number and the percentagday

gnitmethods for RSI included fluoroscopy and ultrasound (totalof 4.2% or 3 cases). Overall, 65% of RSI were abdominal,8% were thoracic, 8% extremity, 2% head/neck, and 10%

immediate (or ,24 hours) removal and proceeding toward .3,000S.P. Stawicki et al. RSI: Need for team training, early recoinvolved a natural orifice (mouth, rectum, vagina, andurinary tract). Table 4 lists RSI locations sorted by commonprocedure/incision types.

Surgical pathology findings

Detailed pathology findings associated with RSI cases inthis study are presented in Table 3. The most common re-ported pathology findings included localized tissue fibrosis(n 5 23), exudative reaction (n 5 22), purulence/abscess(n 5 13), and generic descriptors of inflammatorychange (n 5 11). Findings of fibrosis were significantlymore common in the combined subacute and chronicgroups (n 5 17) compared with immediate and acutegroups (n 5 6, P , .01). The overall proportion of pathol-ogy findings indicative of inflammatory changes and/or thepresence of clinical findings consistent with inflammatorystate (ie, fever, tachycardia, abscess, fistula, other infec-tion) increased significantly over time, with the percentageof inflammation-related findings being 16% in the imme-diate group, 42% in the acute group, 50% in the subacutegroup, and 68% in the chronic group (chi square, P ,.02). In addition, the finding of abscess/mass was muchmore common in the chronic group (58%) than in the other3 temporal groups combined (15%) (Fisher exact test, P ,.01). Moreover, the cumulative incidence of inflammatoryfindings was significantly greater among patients withtextile (ie, surgical sponge) RSI (23 of 42 or 55%) than pa-tients with nontextile (ie, surgical instrument) RSI (10 of 29or 34%, Fisher exact, P , .01).e of retained items retrieved at each temporal marker, starting withs from the index RSI operation.

ion, and prompt retrieval 5Intraoperative error types, safety omissions,and variances

An overview of all recorded SOVs is presented inTable 1. Our previous report suggested that SOV is animportant contributor to overall RSI risk.9 For the presentstudy, we performed a more detailed analysis of SOVfrom the original study that also included 12 additional casesthat occurred outside the study hospitals but that weretreated during the study period at participating centers. Spe-cifically, much greater resolution of data regarding co-occurrence of SOVs and differentiation between individ-ual/team errors is provided in this article. Detailed chart re-view of cases in this series demonstrated that isolated humanerrors (ie, errors attributable to a single individual withoutteam involvement) were seen in only 7 of 71 cases(,10%). Team or system errors, where no single individualcould be pinpointed as an isolated contributor, were seenin 50 of 71 cases (70%). In addition, 14 cases, a combinationof human and team/system errors, occurred. Summarizingthe above findings, more than 90% of RSI events were notedto be associated with some type of team/system error.

A grid displaying co-occurring intraoperative SOVs ispresented in Fig. 1. Fewer than half (34 of 71) RSIs wereattributed to an isolated intraoperative event. Consequently,most (37 of 71) cases involved 2 or more SOVs. Within the

Pelvis 2 14.3Mid-abdomen 2 14.3

Left upper quadrant 1 7.1Other (unspecified) 2 14.3

Laparotomy (n 5 27)

Left upper quadrant 5 18.5Pelvis 5 18.5Right upper quadrant 3 11.1Subcutaneous 3 11.1Subfascial 2 7.41Right lower quadrant 2 7.41Table 4 Locations of identified RSI by the location/case type

Procedure/Item locationNumberof items % Cases

Cesarean section (n 5 14)*Subcutaneous 5 35.7Colic gutter 2 14.3

6latter subset, 18 cases involved co-occurrence of 2 factors,15 cases involved 3 factors, and 4 cases involved 4 or morefactors. The authors acknowledge that because of reportingbiases and medical record inaccuracies, the total numberof contributory factors is likely higher than documentedhere and that our current data represent the best effortat reconstructing the operative events associated with RSIoccurrences.

Comments

This study helps to fill an important void in ourknowledge regarding the natural history of RSIs. For thefirst time, a natural history of RSI clinical presentations,item locations, pathologic findings, and temporal charac-teristics has been clearly elucidated. In addition, wecategorize and describe intraoperative factors associated

Paramedian 2 7.41Retroperitoneal 2 7.41Intraluminal 1 3.70Other (unspecified) 2 7.41

Laparoscopic (n 5 6)

Rectus muscle/sheath 2 33.3Right abdomen (upper andmid-abdomen)

2 33.3

Retroperitoneum 1 16.7Pelvis 1 16.7

Extremity (n 5 6)Subcutaneous 2 33.3Deep wound location 2 33.3Joint space or adjacent to joint 2 33.3

The laparotomy category included both subcostal and midline

incisions. Because of small number of cases in each respective incision

type/location category, thoracic, head/neck, and natural orifice RSI

cases are not listed.

RSI 5 retained surgical item.*All cesarean sections performed via Pfannenstiel incision.Laparotomy and laparoscopy cases combined for different incision

types.with RSI occurrences, shedding some new light on theanatomy of how RSI events may actually take place.

Perhaps, the most significant findings of this studypertain to the detailed forensic examination of intraoper-ative SOVs. Although it intuitively makes sense that itusually takes more than 1 person or 1 event for an RSI tooccur, little published evidence actually supports thisassumption. Our observations that most RSI cases werebecause of team/system errors is consistent with the morerecent recognition that blame does not simply rest with asingle individual in most cases and will continue to requireendorsement of a just culture.10 In addition, our findingsalso provide yet another piece of evidence to validate theswiss cheese model of medical error.9,11 The supportfor this finding comes from the observation that most casesin this study were characterized by co-occurrence of 2 ormore safety variances or omissions. Although intuitivelythe swiss cheese model makes great sense, little empiricalevidence supports this model outside the present study.

Our data show that identification of most RSIs occurs earlyduring the postprocedural hospital course. Most cases in thisseries (62%) were either elective surgeries done on the day ofadmission or emergency cases done early after initial patientpresentation. This distribution of cases is in keeping withpreviously published data for ambulatory, emergency, andsame-day surgery.1215 Our data likely reflect the overall casedistribution patterns across typical tertiary care environments.

The present study reveals clinically relevant datapertaining to RSI anatomic locations (Table 1). Althoughthe overall anatomic distribution of RSI seems somewhatunpredictable, we propose that certain important patternshave emerged. In fact, some locations seem to be morefrequently seen when cases grouped by general type ofcase (ie, cesarean section, laparotomy, laparoscopy, extrem-ity). For example, we found that more than 35% of itemsretained during cesarean sections were located subcutane-ously. These findings support the need for a standardizedreview of the subcutaneous areas before skin closure inevery case as part of the final closing count and surgicalsign out. For the remaining cases, a thorough preclosureabdominal exploration (pelvis, mid-abdomen, and colicgutters) could identify nearly 70% of missed items. Whenreviewing the laparotomy group, simple examination ofbilateral upper quadrants and the pelvis could potentiallyidentify nearly half of all RSIs. Anatomic distribution of re-tained items for laparoscopic procedures likely reflects acombination of case-specific and technique-specific loca-tions. That is to say, one third of retained items in thisgroup were found at the level of the rectus muscle/sheathand an additional third were found on the right side ofthe abdomen. This is likely a reflection of the commonlyperformed laparoscopic appendectomy and cholecystec-tomy being right-sided procedures. RSIs in various ex-tremity locations are equally distributed between the deepspaces, the superficial spaces, and the joint spaces. Com-bined, these types of findings can support the use of hard-

The American Journal of Surgery, Vol -, No -, - 2014wired checklists, whereby the operative sign out would

gnitinclude a regimented, systematic review of the operativeanatomic area.

The authors were not entirely surprised to find out thatnearly 17% of RSIs originated at nonparticipatinginstitutions; this would support the justification for thecreation of a national database that facilitates reporting andtracking of RSIs. Even among tertiary referral centersdedicated to the detection and tracking of RSIs, we canspeculate that as many as 1 in 6 occurrences might beidentified at another institution and may be beyond ourability to study the precipitating risk factors and do acredible root cause analysis to prevent such occurrences. Inaddition to reflecting a small but noticeable trend togravitate away from the original institution among patientswho experience complications,16 this figure also providesus with a better understanding of the magnitude of RSIsthat are not detected or seen at the institution at whichthe index operation occurred. Although similar data onRSI has not been previously published, some inferencescan be drawn from data pertaining to clinical discovery set-tings for other surgical never events and complications. Onestudy has demonstrated that more than 9% of surgical siteinfections were detected either in outpatient setting or atanother hospital.17 Similarly, the occurrence of undocu-mented stage IV pressure ulcers on admission can be, ona scale of severity of medical miss, compared with unin-tentionally leaving a surgical item within the operativefield. The incidence of such undocumented stage IV pres-sure ulcers, despite their obvious appearance and character-istics, has been reported to be as high as 18%,18 which isnot dissimilar to the above RSI data from this study.

It is important to note that the presentation of RSIs tendsto be related to the overall duration of RSI retention period.The initial peak in identification is associated with imme-diate detection, whether acutely symptomatic (usuallybecause of pain) or incidentally found on routine post-operative imaging. The second peak occurs between theinitial 24 hours and 3 months after the unintentional itemretention. This group usually presents with a constellation ofinflammatory-related findings, such as fevers, abscess, fluidcollection, abnormal wound drainage, or fistula.19,20 Finally,patients who present with a long-term delay represent aunique population that remained largely asymptomatic foryears and are either found to have an RSI on incidental im-aging or present with a new unexplained mass that can oftenbe initially suspected as being a tumor.1921 Based on theseobservations, we recommend that plain imaging be consid-ered for any patient with unexplained symptoms after sur-gery to evaluate for RSI, followed by CT imaging only incases where plain imaging is inconclusive.

Most of the RSI that were retrieved within the first24 hours had no significant pathologic or clinical findings.However, the most retained items retrieved beyond theinitial 24-hour period of retention were associated with 1 ormore abnormal pathology finding in conjunction with highlikelihood of clinically significant symptoms. Our observa-

S.P. Stawicki et al. RSI: Need for team training, early recotions of fibrosis, exudative reaction, abscess, and otherinflammatory-related findings and symptoms are in keepingwith the progressive inflammatory process that is frequentlyinitiated by the RSI.20,22 More specifically, the presence ofeither clinical and/or pathologic findings consistent withinflammation rises from 16% in the immediate removalgroup to nearly 70% in the chronic retention group.Furthermore, the incidence of abscess/mass was nearly 4times higher in the chronic group than in the 3 other groupscombined. Not unexpectedly, textile RSIs were also associ-ated with greater incidence of inflammatory-related clinicalsymptoms and signs, although this did not translate intoa greater overall proportion of symptomatic cases. Lesscommon were chronic manifestations of RSIs, includinggranuloma formation, tissue erosion, mass, and fistula/sinusformation. These findings are generally consistent with pre-viously published expert opinions and case-based descrip-tive reports.1,20,21,2325 Consequently, early identificationand removal of RSI are desirable to avoid long-termsequelae and potentially reduce the need for more invasiveand complex interventions that may be required later in thenatural course of RSI retention. It must be emphasized thatnearly half of the patients in our series had more than 1symptomatic complaint. Such data have not previouslybeen tabulated, and all inferences regarding the natural his-tory of RSI have been based on expert opinions, case series,isolated case reports, and reviews.1,20,21,2325

Limitations of this study include its retrospective nature,relatively small patient sample size, and lack of uniform dataavailability for all variables across all reported cases. Thecategorization of clinical signs/symptoms and pathologicfindings into narrowly defined groups limits the overallresolution of the reported data. Also regarding symptomsand clinical sequelae of RSI, the optimal study wouldinclude a survey that directly solicits patient informationregarding long-term quality-of-life and symptoms. However,because of the sensitive nature of this investigative area, andour studys contractual restrictions, such long-term follow-up is not feasible. In addition, intraoperative events reportedhere are limited by the availability and the inherent biasof retrospectively recorded documentation. Thus, it is verylikely that conclusions related to co-occurrence of more than1 SOV in our study sample could have been strengthenedfurther by more accurate, prospective recording of SOVs.Similarly, there is likely some degree of reporting bias,especially for more subjective variables (ie, those pertainingto imaging/tagging interpretation, team communication, oradherence to established protocols). Finally, we do not havea true denominator of SOVs that do not result in RSI.Consequently, our findings should be viewed primarily as apurely descriptive attempt in understanding this multifacetedpatient safety problem.

Conclusions

The finding that most operations complicated by RSIswere found to involve team/system errors and more than 2

ion, and prompt retrieval 7SOVs support the need for institutional emphasis on team

training. The observation that early RSI removal minimizespatient morbidity and symptoms highlights the need forprompt RSI identification and treatment, especially in lightof our finding that the incidence of inflammation-relatedpathology and clinical findings increase significantly withlonger periods of surgical item retention.

References

1. Stawicki SP, Evans DC, Cipolla J, et al. Retained surgical foreign

bodies: a comprehensive review of risks and preventive strategies.

Scand J Surg 2009;98:817.

2. Yamamura N, Nakajima K, Takahashi T, et al. Intra-abdominal texti-

loma. A retained surgical sponge mimicking a gastric gastrointestinal

stromal tumor: report of a case. Surg Today 2008;38:5524.

3. Wang CF, Cook CH, Whitmill ML, et al. Risk factors for retained

surgical foreign bodies: a meta-analysis. OPUS 12 Scientist 2009;3:

217.

4. Yeung KW, Chang MS, Huang JF. Imaging of transmural migration of

a retained surgical sponge: a case report. Kaohsiung J Med Sci 2004;

20:56771.

5. Manikyam SR, Gupta V, Gupta R, et al. Retained surgical sponge pre-

senting as a gastric outlet obstruction and duodeno-ileo-colic fistula:

report of a case. Surg Today 2002;32:4268.

6. Kato K, Kawai T, Suzuki K, et al. Migration of surgical sponge re-

tained at transvaginal hysterectomy into the bladder: a case report.

Hinyokika Kiyo 1998;44:1835.

10. Healy GB, Barker J, Madonna G. Error reduction through team leader-

ship: seven principles of CRM applied to surgery. Bull Am Coll Surg

2006;91:246.

11. Reason J. Human error:models andmanagement. BMJ 2000;320:76870.

12. Mochizuki T, Takehara Y, Ichijo K, et al. Case report: MR appearance

of a retained surgical sponge. Clin Radiol 1992;46:667.

13. Moffatt-Bruce SD, Ellison EC, Anderson HL 3rd , et al. Intravascular

retained surgical items: a multicenter study of risk factors. J Surg Res

2012;178:51923.

14. Matsuki M, Matsuo M, Okada N. Case report: MR findings of a re-

tained surgical sponge. Radiat Med 1998;16:657.

15. Childers JM, Caplinger P. Laparoscopic retrieval of a retained surgical

sponge: a case report. Surg Laparosc Endosc 1993;3:1358.

16. Abdul-Karim FW, Benevenia J, Pathria MN, et al. Case report 736: re-

tained surgical sponge (gossypiboma) with a foreign body reaction and

remote and organizing hematoma. Skeletal Radiol 1992;21:4669.

17. al-Salem AH, Khwaja S. Intestinal obstruction due to retained, eroding

surgical sponge. Case report. Acta Chir Scand 1989;155:199200.

18. Sexton CC, Lawson JP, Yesner R. Case report 174: cottonballoma of

femur (due to retained surgical sponge with foreign body giant cell re-

action). Skeletal Radiol 1981;7:2113.

19. Rappaport W, Haynes K. The retained surgical sponge following intra-

abdominal surgery. A continuing problem. Arch Surg 1990;125:4057.

20. Dembitzer A, Lai EJ. Images in clinical medicine. Retained surgical

instrument. N Engl J Med 2003;348:228.

21. Massimiliano PA, Massimo PS. Retained intra-abdominal surgical in-

strument: a rare condition of acute abdomen. ANZ J Surg 2010;80:758.

22. Guner A, Hos G, Kahraman I, et al. Transabdominal migration of

retained surgical sponge. Case Rep Med 2012;2012:249859.

23. Reece M, Troeleman ND, McGowan JE, et al. Reducing the incidence

8 The American Journal of Surgery, Vol -, No -, - 2014in the thigh: report of the third known case in the limb. Radiat Med

2003;21:2202.

8. Stawicki SP, Moffatt-Bruce SD, Ahmed HM, et al. Retained surgical

items: a problem yet to be solved. J Am Coll Surg 2013;216:1522.

9. Reason J. Safety in the operating theatredpart 2: human error andorganisational failure. Qual Saf Health Care 2005;14:5660.of retained surgical instrument fragments. AORN J 2011;94:3014.

24. Karahasanoglu T, Unal E, Memisoglu K, et al. Laparoscopic removal

of a retained surgical instrument. J Laparoendosc Adv Surg Tech A

2004;14:2413.

25. Reason J. Human error: models and management. West J Med 2000;

172:3936.7. Kominami M, Fujikawa A, Tamura T, et al. Retained surgical sponge

Natural history of retained surgical items supports the need for team training, early recognition, and prompt retrievalMethodsResultsDescriptive characteristics of the retained surgical item sampleTiming of retained surgical item procedures and retained surgical item identification/retrievalClinical signs and symptomsDiagnostic imaging and retained surgical item locationSurgical pathology findingsIntraoperative error types, safety omissions, and variances

CommentsConclusionsReferences