An Investigation of Bloodborne Pathogen Transmission Due to Multipatient Sharing of Insulin Pens

Transcription

1 MILITARY MEDICINE, 177, 8:930, 2012 An Investigation of Bloodborne Pathogen Transmission Due to Multipatient Sharing of Insulin Pens Shilpa Hakre, DrPH, MPH*; LTC Donna R. Upshaw-Combs, MC USA ; Eric E. Sanders-Buell, BS ; Stephanie L. Scoville, DrPH ; CPT Joshua D. Kuper, MC USA ; Linda L. Jagodzinski, PhD ; Andrea N. Bradfield, BS ; Dinae C. Davison, RHIT ; CPT William G. Callis, MC USA ; Angela B. Owens, MPH**; COL Nelson L. Michael, MC USA ; LTC Robert J. O Connell, MC USA ; Sheila A. Peel, MSPH, PhD ; COL John W. Gardner, MC USA (Ret.) ; Nicola D. Thompson, PhD ; Dale J. Hu, MD, MPH ; COL Jerome H. Kim, MC USA ; Sodsai Tovanabutra, PhD ; Paul T. Scott, MD, MPH ; LTC(P) Sandra G. LaFon, MC USA ; the Insulin Pen Investigation Team ABSTRACT On January 30, 2009, nursing staff at a military hospital in Texas reported that single-patient use insulin pens were used on multiple patients. An investigation was initiated to determine if patient-to-patient bloodborne transmission occurred from the practice. Human immunodeficiency virus (HIV), hepatitis B virus (HBV), and hepatitis C virus (HCV) testing was offered to patients hospitalized from August 2007 to January 2009 and prescribed insulin pen injections. Virus from HCV-infected patients sera was sequenced and compared for relatedness. An anonymous survey was administered to nurses. Of 2,113 patients prescribed insulin pen injections, 1,501 (71%) underwent testing; 6 (0.4%) were HIV positive, 6 (0.4%) were hepatitis B surface antigen positive, and 56 (3.7%) had HCV antibody. No viral sequences from 10 of 28 patients with newly diagnosed and 12 of 28 patients with preexisting HCV infection were closely related. Of 54 nurses surveyed, 74% reported being trained on insulin pen use, but 24% believed nurses used insulin pens on more than one patient. We found no clear evidence of bloodborne pathogen transmission. Training of hospital staff on correct use of insulin pens should be prioritized and their practices evaluated. Insulin pens should be more clearly labeled for single-patient use. *Epidemiology and Threat Assessment, U.S. Military HIV Research Program, Henry M. Jackson Foundation for the Advancement of Military Medicine, 6720-A Rockledge Drive, Suite 400, Bethesda, MD William Beaumont Army Medical Center, 5005 North Piedras Street, El Paso, TX Molecular Virology and Pathogenesis, U.S. Military HIV Research Program, Henry M. Jackson Foundation for the Advancement of Military Medicine, 503 Robert Grant Avenue, Silver Spring, MD Preventive Medicine Residency Program, U.S. Army Public Health Command, 503 Robert Grant Avenue, Silver Spring, MD khiv Diagnostics and Reference Laboratory, U.S. Military HIV Research Program, Walter Reed Army Institute of Research, 13 Taft Court, Suite 101, Rockville, MD Moncrief Army Community Hospital, 4500 Stuart Street, Fort Jackson, SC **Ke aki Technologies, LLC, U.S. Army Public Health Command, Public Health Region South, 2472 Schofield Road, 2D FL, Fort Sam Houston, TX U.S. Military HIV Research Program, Walter Reed Army Institute of Research, 6720-A Rockledge Drive, Suite 400, Bethesda, MD Uniformed Services University of the Health Sciences, 4301 Jones Bridge Road, Bethesda, MD Centers for Disease Control and Prevention, 1600 Clifton Road, MS G-37, Atlanta, GA. kk1st Armored Division, Sergeant Major Boulevard, Fort Bliss, El Paso, TX This work was presented in part (by invitation) at the Centers for Disease Control and Prevention Sticking with Safety: Eliminating Bloodborne Pathogen Risks During Blood Glucose Monitoring meeting, May 3, 2010, Atlanta, Georgia, and at the 13th Annual Force Health Protection Conference, August 11, 2010, Phoenix, Arizona. The views expressed are those of the authors and should not be construed to represent the positions of the U.S. Department of Defense or the Centers for Disease Control and Prevention. INTRODUCTION Reports of acute viral hepatitis infections to public health authorities and subsequent public health investigations often reveal the cause as failure to follow infection control or manufacturer guidelines, which have resulted in patient-to-patient virus transmission in health care settings. 1 5 Some examples of infection control breaches implicated in hepatitis C virus (HCV) and hepatitis B virus (HBV) transmissions among patients include reusing syringes and single-use vials on multiple patients at endoscopy clinics, 1,5 reusing syringes and saline bags on multiple patients at an oncology clinic, 2 and sharing of blood glucose monitoring devices A recent survey that assessed adherence to infection control practices among 68 Medicare-reimbursed ambulatory surgical centers in three states revealed more than half were found to have at least 1 infection control breach. A commonly identified breach was failure to handle blood glucose monitoring equipment appropriately. 11 The insulin delivery pen or insulin pen was introduced in the United States in 1987 and is a single fountain pen-like device, which replaces the traditional syringe, vial, and needle method of insulin delivery into a single unit. 12 The pen is manufactured either as a reusable device for use in a single patient with a new needle for each injection or a disposable device for single use. Purported advantages of an insulin pen over the vial/syringe system of insulin administration include accuracy in dosage, increased patient adherence, safety, and convenience. On January 30, 2009, the nursing staff at a military hospital in Texas reported the incorrect use of insulin pens to their 930

2 superiors and corrective actions were taken. Insulin pens reportedly were used on more than one patient, with a new sterile disposable needle used for each injection. The insulin pens had been introduced at the facility in August Although no patients newly diagnosed with bloodborne pathogen (BBP) infections were reported from improper use of the pens, since biological material and blood can backflow into the insulin cartridge portion of the pen system, the risk of infectious material passing from patient-to-patient existed. The hospital immediately took corrective actions which prompted several alerts, for example, by the U.S. Food and Drug Administration and the Centers for Disease Control and Prevention to the medical community. 16,17 To date, no report of patient-to-patient BBP transmission via insulin pens has been published although reuse of the pens was reported at a medical center in New York. 18 We describe the results of the ensuing investigation assessing whether patient-to-patient transmission of the three pathogens of primary concern during health care-related blood exposures, human immunodeficiency virus (HIV), HBV, and HCV, occurred among potentially exposed hospitalized patients METHODS Patient Notification and Testing The medical staff at the 137-bed hospital attempted to notify all living patients prescribed an insulin pen injection while admitted from August 1, 2007 to January 30, 2009 at the facility and offered testing and medical care if found infected. The hospital alerted the public of the potentially harmful practice through press releases to the news media and encouraged patients to call a toll-free helpline for information, patient concerns, and scheduling of testing and care. Testing was offered to patients at initial notification and at follow-up 6 months after the date of their last insulin order. Patients who screened negative on testing were notified of their results by certified mail. Patients with test results suggesting a BBP infection were notified verbally and referred for specialty care and treatment. Case Definitions We considered patients as case if their laboratory test results indicated serologic evidence of HIV, HBV, or HCV or if a diagnosisofhiv,hbv,orhcvonelectronicmedicalor death records was confirmed by serologic evidence on historical laboratory records. Patient serum samples were sent to contract laboratories for screening and supplemental confirmatory testing for HIV, HBV, and HCV. Samples were screened for antibody to HIV (anti-hiv) at ViroMed Laboratories (Minnetonka, Minnesota). Screening for total and immunoglobulin M antibodies to hepatitis B core antigen (anti-hbc), hepatitis B surface antigen (HBsAg), antibodies to HCV (anti-hcv), and clinical viral genotype testing (hepatitis C viral RNA genotype, line probe assay [LiPA]) for patients with HCV RNA were performed at Quest Diagnostics (Irving, Texas). Samples that screened positive underwent supplemental confirmatory tests: HIV western blot (WB) and HIV RNA if indeterminate for WB; HBV neutralization assays and HBV DNA viral load if HBsAg confirmed positive; HCV recombinant immunoblot assay (RIBA), HCV RNA if RIBA was indeterminate. 22 A case of chronic HBV infection was a patient who confirmed positive both for HBsAg and total anti-hbc or confirmed positive for HBsAg with detectable HBV DNA. Patients were considered to have resolved HBV infection if they were anti-hbc positive and HBsAg negative. Case patients were classified further as newly diagnosed or preexisting. Newly diagnosed cases had no prior awareness of infection at notification of test results, and electronic medical and laboratory records revealed no indication of infection before their initial insulin pen prescription in the period under investigation. Preexisting cases were patients with knowledge of their infection at notification of test results and before their initial insulin pen prescription in the period under investigation, or historical electronic medical and laboratory records revealed infection before initial insulin pen exposure. Epidemiologic Links A timeline of insulin pen exposures was constructed from admission, insulin pen order, and discharge dates for all patients with laboratory results. To identify opportunities for patientto-patient virus transmission via common insulin pen use, an epidemiological link was defined as a newly diagnosed case sharing temporal (i.e., on the same or within 10 days of) and location (i.e., on the same hospital floor) overlap with a preexisting case having the same infection during a hospital stay. Virus Characterization Infected patients were asked to provide blood samples for virus subtyping and sequencing to assess the relatedness of virus between patients and to further evaluate potential patient-topatient virus transmission. For HCV characterization, partial genome sequencing and analysis of the nonstructural 5 b (NS5B) and core envelope 1 (C/E1) regions as described by Murphy et al (2007) 23 and investigation of sequence relatedness using signature analysis were performed on cryopreserved sera with detectable viral load. Nurse survey Anonymous surveys were administered to all available nurses on each floor (including floor C where improper pen use was first identified) during a single shift change. The survey was designed to subjectively assess how common the practice of insulin pen reuse was among all nurses and did not ask about personal experience. Data Sources Information management staff identified all hospitalized patients with an insulin pen prescription during the potential exposure 931

3 period, the at-risk cohort, from the hospital s pharmacy database. Electronic administrative (Defense Finance and Accounting Service, Social Security), medical, and laboratory records (the Department of Defense Armed Forces Health Longitudinal Technology Application electronic health record system, Veterans Affairs [VA]) supplied current patient contact information, prior medical history of HIV, HBV, and HCV, and demographic information. Up-to-date longitudinal electronic medical records (Defense Medical Surveillance System) 24 for the at-risk cohort provided hospitalization admission and discharge dates, demographic information, and pre-insulin pen exposure BBP medical diagnoses (from International Classification of Disease, 9th Revision codes). Electronic death index records (Texas Department of Health Services Vital Statistics Unit) determined causes of death for deceased patients in the cohort. Data Analysis The prevalence of HIV, HBV, and HCV infection was calculated among patients tested in the at-risk cohort. For each patient, total patient days of potential insulin pen exposure were calculated by summing the interval between the first date insulin pen injections were prescribed during an admission and discharge date for that admission. The rate of epidemiologic links on each floor was compared to other floors by calculating the proportion of epidemiologic links on each floor and comparing the proportion to a referent floor; proportions were calculated by dividing the number of links by the total number of admissions having insulin pen prescriptions. Chi-square, Student s t-test, and Kruskal Wallis tests compared statistical differences in patient groups and two-sided p-values were assessed at 95% significance level. Data management and statistical analysis were conducted using Epi Info version (CDC, Atlanta, Georgia), Statistical Analysis Software version 9.1 (SAS Institute, Cary, North Carolina), and Programs for Epidemiologists (WINPEPI) version RESULTS Patient Notification and Testing From August 1, 2007 to January 30, 2009, 2,113 hospitalized patients were prescribed injections from insulin pens. The cohort of 2,113 patients were predominantly older (71% ³ 55 years), Veteran (33%) or retired (27%), and male (68%) (Table I). During the period at risk, 2,092 patients had 3,804 admissions totaling 27,023 inpatient days, of which 15,160 (56%) were potential insulin pen exposure days; admission and discharge dates were missing for 21 patients in the cohort. Average potential duration of insulin pen exposure was 7.2 days (median = 4.0) (Table I). Among the 2,113 patients in the at-risk cohort, 1,501 (71%) participated in testing for HIV, HBV, and HCV infection. Patients not tested (n = 612) were not significantly different from those tested by TABLE I. Demographic and Exposure Characteristics of 2,113 Patients Prescribed Insulin Pen Injections, August 2007 to January 2009 Characteristic Cohort (n = 2,113) Tested (n = 1,501) Not Tested (n = 612) p-value Age a < Mean (Median) 62.2 (63.7) 61.2 (62.8) 64.8 (67.3) (28) 441 (29) 156 (26) (45) 725 (48) 220 (37) (26) 335 (22) 220 (37) Gender 0.12 Female 685 (32) 502 (33) 183 (30) Male 1428 (68) 999 (67) 429 (70) Race 0.08 White 940 (49) 680 (50) 260 (43) Other b 718 (34) 499 (37) 219 (36) Black 242 (11) 56 (10) 56 (9) Unknown 213 (10) 136 (9) 76 (12) Military Beneficiary Category < Veterans 702 (33) 469 (31) 233 (38) Retired Military 570 (27) 439 (29) 131 (21) Beneficiary/Dependent 585 (28) 441 (29) 144 (24) Other c 256 (12) 152 (10) 103 (17) No. of Admissions With Insulin Pen Prescriptions d 0.24 Mean (Median) 1.4 (1.0) 1.4 (1.0) 1.3 (1.0) (78) 1158 (78) 479 (80) 2 or More (Maximum = 11) 455 (22) 334 (22) 121 (20) Length of Exposure, Total, Days < Mean (Median) 7.2 (4.0) 6.5 (3.0) 9.1 (5.0) <3 791 (37) 599 (40) 192 (31) (33) 510 (34) 183 (30) 7 or More (Maximum = 208) 629 (30) 392 (26) 237 (39) n (%) unless indicated otherwise. a Date of birth was missing for 16 patients. b Other: Hispanic, American Indian/Pacific Islander, other. c Other: civilian humanitarian, active duty, National Guard, deceased sponsor, and 1 unknown. d Admission dates were unknown for 21 patients. 932

4 FIGURE 1. Follow-up of patients prescribed insulin pen injections at the hospital where improper insulin pen use was identified, August 2007 to January *Patients prescribed insulin pen injections during admissions from August 1, 2007 to January 30, 2009; an additional 21 patients who were not part of the 2,113 at-risk cohort also requested and completed testing (none of these patients tested positive for HIV, HBV, or HCV). An additional 157 patients were possibly exposed to HBV (anti-hbc positive, HBsAg negative). Electronic medical or death index records; 25 patients had historical laboratory records that confirmed diagnosis indicated on medical and death records; newly diagnosed patients were those who first tested positive after the exposure period per laboratory records. race or gender but differed significantly by older age, longer potential duration of exposure to insulin pens, and health care eligibility (military beneficiary category) (Table I). TABLE II. Among 1,501 patients tested, 68 (4.5%) cases were identified; a majority(n = 56, 82%) tested anti-hcv positive (Fig. 1). All 6 cases with HIV infection were preexisting, as were 3 of Anti-HCV and Total Anti-HBc Prevalence by Demographic Characteristics and Period of Diagnosis Anti-HCV Total Anti-HBc Characteristic Total Number Tested Number Positive Prevalence (%) 95% CI Number Positive Prevalence (%) 95% CI Age Gender Female Male Race White Other Black Unknown Military Beneficiary Category Veterans Retired Military Beneficiary/Dependent Other Diagnosis Before Pen Exposure Yes No

5 6 cases with chronic HBV infection. One hundred fifty-seven patients had resolved HBV infection. No patients were coinfected with HIV, HBV (HBsAg), or HCV. Among 612 patients not tested, electronic medical and death records review identified 39 patients with BBP diagnoses. Among them, 25 patients had laboratory records indicating positive test results of whom a majority (84%) had preexisting infection (Fig. 1). FIGURE 2. Potential opportunities for transmission via common use of insulin pens among HCV-infected patients with overlapping hospital stay at the medical facility where multipatient sharing of insulin pen was identified, August 2007 to January The schematic illustrates a line list of 52 HCVinfected patients identified in the investigation who were prescribed insulin pen injections and who had overlap in hospital stay on each floor of the medical facility during the period when insulin pen cartridges may have been shared. Patient identification numbers in parentheses are patients with preexisting HCV infection who may have been potential sources of HCV infection and epidemiologically linked with newly diagnosed HCV infected patients. Days indicated on the line list are the admission and 10 days postdischarge dates for each patient and are shaded as dark blue for preexisting HCV-infected patients and as light blue for newly diagnosed HCV-infected patients. The three floors of the medical facility are indicated by the letters A (intensive care, units A1 to A3), B (surgical care), and C (medical care). The HCV genotype for patients is indicated when available. The reasons for missing genotype are indicated as ND, RNA not detected; NT, patient not tested; NS, no sample provided for genotype determination. *Genotype was determined by sequencing. 934

6 Prevalence Estimates Among patients tested, anti-hcv prevalence was 3.7% and was highest among those aged 15 to 54 years (5.4%), males (4.6%), and Veterans (7.7%) (Table II). Of those tested, 0.4% patients tested anti-hiv positive and 0.4% had chronic HBV infection (HbsAg and anti-hbc positive). Among 163 (10.8%) patients with past exposure to HBV (anti-hbc positive), prevalence was highest among patients aged 75 to 95 years (16.4%), males (11.8%), and retirees (14.6%) (Table II). Epidemiological Links Thirty-four potential epidemiologic links were identified among 24 newly diagnosed cases and 28 preexisting cases (Fig. 2). Although the frequency of epidemiological links varied by floor and by patient, none of the rates of epidemiological links compared to any one floor achieved statistical significance. Although Floor C had the highest frequency of epidemiological links (17 epi links/1298 admissions), Floor A had the highest rate (14 epi links/978 admissions; rate ratio [RR] = 1.9) in comparison to Floor B (4 epi links/535 admissions; RR = 1.0) and all three floors (35 epi links/2811 admissions; RR = 1.1). None of the relative rates of epidemiological links compared to Floor B achieved statistical significance (95% confidence interval [CI]: Floor B referent, ; all floors referent, ). Virus Characterization Since no epidemiologic evidence supporting patient-topatient transmission of HIV or HBV was identified, sequencing of HIV or HBV was not pursued. Among 56 patients who tested positive for anti-hcv, 25 (45%) patients had detectable RNA. Of the 56 patients, 39 patients (68%) provided a serum sample for viral sequencing; 21 (54%) were newly diagnosed and 18 were preexisting cases. Of 39 samples, 22 (56%) had detectable RNA that were successfully amplified and sequenced in the NS5B region; 10 were newly diagnosed cases and 12 were preexisting cases. Additionally, 19 of the 22 were amplified and sequenced in the C/E1 region. Sequence analysis of both regions revealed diverse genotypes: 12 were 1a, 7 were 1b, and the remaining 3 were 2a, 2b, and 3a. Among 38 HCV cases that were tested and had shared epidemiological links with a preexisting case, 12 patient samples (32%) were obtained for sequencing, 6 from newly diagnosed and 6 from preexisting cases were sequenced; none were closely related. Similarly, signature analyses of the sequences for all 22 patients with detectable RNA indicated that no sequences from any patient sample were closely related to each other, which excluded transmission between newly diagnosed cases. Among 17 HCV case patients who did not provide a sample for viral sequencing, 3 had detectable RNA, 3 had undetectable RNA, and 11 did not participate in genotype and RNA testing. Of the patients with detectable RNA, 2 were clinically genotyped 1a or 1b and 1 did not complete genotype testing. Nurse Survey An anonymous survey administered to and completed by 54 nurses from three floors indicated that one-quarter (26%) of nurses were not trained for use of the insulin pen (Table III). Furthermore, only 9 (22%) were instructed about the potential for aspiration of blood back into the insulin cartridge. Of those trained, all but one recalled receiving instruction that the pens were specifically for single-patient use. Nurses on Floor C most frequently reported the availability of insulin pens in an open area of the floor s medication room (77%), the occurrence of least one nurse carrying insulin pens in pockets (76%), using a pen not labeled with a patient s name (65%), carrying pens home (53%), and using pens on more than one patient (41%). Nurses reported some of the following possible reasons for misuse of the pens: incorrect instruction, avoidance of delay in administration of insulin, failure to recognize the possibility for biological contamination of the insulin solution, and unavailability of pens from pharmacy. DISCUSSION This investigation found no definitive evidence of virus transmission among patients prescribed insulin pens although patient-to-patient transmission of HIV, HBV, and HCV infection could not be excluded. Even with the use of considerable personnel, financial, and laboratory resources, several features of this investigation, common to retrospective investigations, 2,8,21 may have limited the possibility of documenting patient-to-patient transmission: (1) more than a quarter (29%) of the at-risk cohort was not tested; (2) it was difficult to establish when the inappropriate use of the pens had begun since documentation of insulin administration was not reliable and an interval of up to 18 months had passed since introduction of the pens among hospitalized patients and reporting of the inappropriate use of the pens; (3) among patients with HCV antibody, more than half (55%) either did not have RNA testing or did not have detectable RNA that could be sequenced. Combined, these factors reduced the likelihood of finding relatedness among patient viruses. Although transmission could not be excluded, evidence did not point to a large outbreak because (1) no evidence was found of onset of acute disease after potential exposure to shared insulin pens; (2) sequencing of virus from patient samples indicated no HCV were closely related; (3) the prevalence of infection in this population (>50% preexisting) was comparable to similar populations. The anti-hcv prevalence of 3.7% overall and 7.7% among Veterans in our investigation was comparable to other studies conducted among VA patients. A population-based serosurvey of 20 VA medical centers and HCV screening and referral programs at VA medical centers in Minneapolis and Providence reported anti-hcv prevalence rates of 4%, 5.4%, and 7.3%, respectively Although these VA studies sampled 935

7 TABLE III. Summary of Anonymous Survey Responses by 54 Nurses From Three Floors at the Military Hospital Where Improper Insulin Pen Use was Identified, August 2007 to January 2009 Question Floor A (n = 19) Floor B (n = 18) Floor C (n = 17) All 3 Floors (n = 54) At Military Hospital Before August (37) 4 (22) 3 (18) 14 (26) Pen Experience Before Working at the Military Hospital 8 (42) 4 (22) 2 (12) 14 (26) Military Hospital s Standard Policy and Procedure 12 (63) 10 (59) 6 (35) 28 (53) for Pen Use Available a Instructions on How to Use Pens Received at Military Hospital a 16 (84) 15 (83) 9 (53) 40 (74) Manufacturer Training Supervisor Training Preceptor Training Colleague Training Described Potential for Aspiration of Blood Back into Insulin Cartridge Instructed That Pens Were for Single-Patient Use Assessment of Competency Received Before First Use of Pen b 9 (47) 6 (33) 3 (18) 18 (33) Training on Pen Use Received Outside of Military Hospital 8 (42) 5 (28) 3 (18) 16 (30) Other Employer Nursing School No Response Locations Insulin Pens Were Found Before Patient Use a Pharmacy 12 (63) 13 (72) 9 (53) 34 (63) Patient-specific Bin in Omnicell 14 (74) 17 (94) 14 (82) 45 (83) Open Area in Medication Room 3 (16) 5 (28) 13 (77) c 21 (39) Portable Med Cart 3 (16) 11 (61) 10 (59) 24 (44) Bedside Drawer 4 (21) 1 (6) 0 (0) 5 (9) Subjective Number of Nurses Using a Single Insulin Pen on More Than One Patient a None 14 (74) 17 (94) 10 (59) 41 (76) <5 2 (11) 1 (6) 4 (24) c 7 (13) >5 3 (16) 0 (0) 3 (18) c 6 (11) Subjective Number of Nurses Carrying Insulin Pens on Person (e.g., Pocket) a None 17 (90) 8 (44) 4 (24) 29 (54) <5 1 (5) 7 (39) 4 (24) c 12 (22) >5 1 (5) 3 (17) 9 (53) c 13 (24) Subjective Number of Nurses Carrying Insulin Pens Home a None 17 (94) 18 (100) 8 (47) 43 (81) <5 1 (6) 0 (0) 8 (47) c 9 (17) >5 0 (0) 0 (0) 1 (6) c 1 (2) Subjective Number of Nurses Using Insulin Pen Not Labeled With a Patient s Name a None 16 (84) 18 (100) 6 (35) 40 (74) <5 2 (11) 0 (0) 8 (47) c 10 (19) >5 1 (5) 0 (0) 3 (18) c 4 (7) n (%); Floor A = intensive care, Floor B = surgical care, and Floor C = medical care where the insulin pen sharing was first identified. a Before February b Between August 2007 and January c Availability of insulin pens in an open area of the medication room on Floor C compared to other floors, p < 0.001; subjective reports on Floor C of 1 or more nurses compared to none carrying insulin pens in pockets, p < 0.001, using a pen not labeled with a patient s name, p < 0.001, carrying pens home, p < 0.001, using pens on more than 1 patient, p = populations with varying prevalence, the overall anti-hcv prevalence was similar in that it was higher than the general U.S. population anti-hcv prevalence of 1.6%. 29 Our cohort had a higher prevalence of prior HBV infection (anti-hbc positive) of 10.8%, compared to the national prevalence of 4.8%, as determined by the National Health and Nutrition Examination Survey (NHANES) study, which includes persons aged 6 years and older. 30 Prevalence estimates of HBV infection (those anti-hbc positive) from the NHANES were 7.7% for those aged 50 years and over. 31 The prevalence of anti-hbc positivity is known to rise with increasing age. 30,32 Almost one-fourth of patients tested were over 75 years of age; therefore, it is likely the higher prevalence of resolved HBV infection in our cohort is attributable to their older age. It is probable that most of the HCV-infected patients who were newly diagnosed by the investigation acquired their infection before exposure to the insulin pens. Diagnosis of HCV is often delayed for many years since up to 70% or 80% of patients have no noticeable symptoms that are likely to prompt medical visits and testing. 33 In a cross-sectional serosurvey of Veterans from 20 VA medical centers, 46% of seropositive patients were unaware of having HCV infection

8 The frequency of the pen-sharing practice by nurses during the period under investigation was unknown. Although sharing of the pen was reported by 24% of nurses who believed other nurses used the pens on more than one patient, the practice may have been infrequent. The survey results indicated misuse of pens may be prevented in the future by emphasizing ongoing training of staff in the correct use of insulin pens, periodic evaluations of staffs competency in using insulin pens correctly, and better labeling of pens for single-patient use. The hospital simultaneously had implemented the corrective measures indicated by the nurses on the survey. Despite these limitations and the financial, legal, logistical, and psychological challenges of patient disclosure encountered in an infection control breach, 34 in the absence of evidence of pathogen transmission and unknown or low risk of BBP transmission, the hospital s decision that patient disclosure was the ethical course of action in such an event and the approach of this investigation highlight the importance of patient notification and testing and supports a qualitative approach to management of infection control breaches outlined by Patel et al. 35 Several important outcomes of the investigation included discarding all insulin pens in use when the breach was reported; continuing relabeling of pens and face-to-face staff training on the correct use of insulin pens; an assessment of practices at all U.S. Army medical facilities using insulin pens; 36 identification of patients previously unaware of their infection and, where possible, prevention of secondary transmission to their contacts; medical care and treatment for previously infected patients not in care; and standardization of documentation of insulin administration. In summary, although the possibility of isolated patientto-patient transmission cannot be excluded, the prevalence rates of BBPs among persons tested indicate rates and risk factors of newly diagnosed patients were comparable to population-based studies among Veterans. Knowledge deficit in infection control risks were common. Prioritization of ongoing and improved training of staff before implementation and use of insulin pens, ongoing assessment of staff to assess adherence to manufacturers guidelines, and improved device labeling and instructions for use by manufacturers will likely help to reduce the potential risks for patients and prevent considerable psychological, legal, financial, and logistical impact on patients and medical institutions as a result of errors in insulin pen use and the disclosure of these errors. ACKNOWLEDGMENTS This investigation would not have been possible without the patient contact and notification efforts of the Insulin Pen Investigation Team members, Mr. Bruce Gramlich, CPT Billy McPherson, CPT Michael Swanhart, MAJ Victoria Prehn, and CPT Alfreda Ritter. We thank MAJ Christopher Perdue at the Armed Forces Health Surveillance Center for his assistance in obtaining archived medical encounter records for all patients in our cohort. We are grateful for laboratory assistance from Ms. Ying Liu and sample processing efforts of Mr. J. Connor Eggleston (U.S. Military HIV Research Program). We thank Mrs. Faye Raye at the Texas Department of Health Services Vital Statistics Fraud Unit for extracting cause of death records. We thank Drs. Philip Spradling and Joseph Perz for their review of the manuscript and Dr. Yuri Khudyakov for his review of the HCV molecular characterization results from the investigation (Centers for Disease Control and Prevention). The investigation, part of the response to the discovery of the insulin pen misuse, was funded by the William Beaumont Army Medical Center, the Henry M. Jackson Foundation, and the U.S. Army Public Health Command. REFERENCES 1. Gutelius B, Perz JF, Parker MM, et al: Multiple clusters of hepatitis virus infections associated with anesthesia for outpatient endoscopy procedures. Gastroenterology 2010; 139(1): Macedo de Oliveira A, White KL, Leschinsky DP, et al: An outbreak of hepatitis C virus infections among outpatients at a hematology/oncology clinic. Ann Intern Med 2005; 142(11): Comstock RD, Mallonee S, Fox JL, et al: A large nosocomial outbreak of hepatitis C and hepatitis B among patients receiving pain remediation treatments. Infect Control Hosp Epidemiol 2004; 25(7): Germain JM, Carbonne A, Thiers V, et al: Patient-to-patient transmission of hepatitis C virus through the use of multidose vials during general anesthesia. Infect Control Hosp Epidemiol 2005; 26(9): Centers for Disease Control and Prevention: Acute hepatitis C virus infections attributed to unsafe injection practices at an endoscopy clinic Nevada, MMWR Morb Mortal Wkly Rep 2008; 57(19): Counard CA, Perz JF, Linchangco PC, Christiansen D, et al: Acute hepatitis B outbreaks related to fingerstick blood glucose monitoring in two assisted living facilities. J Am Geriatr Soc 2010; 58(2): Dreesman JM, Baillot A, Hamschmidt L, Monazahian M, Wend UC, Gerlich WH: Outbreak of hepatitis B in a nursing home associated with capillary blood sampling. Epidemiol Infect 2006; 134(5): Desenclos JC, Bourdiol-Razes M, Rolin B, et al: Hepatitis C in a ward for cystic fibrosis and diabetic patients: possible transmission by springloaded finger-stick devices for self-monitoring of capillary blood glucose. Infect Control Hosp Epidemiol 2001; 22(11): Centers for Disease Control and Prevention: Transmission of hepatitis B virus among persons undergoing blood glucose monitoring in long-term-care facilities Mississippi, North Carolina, and Los Angeles County, California, MMWR Morb Mortal Wkly Rep 2005; 54(9): Thompson ND, Perz JF: Eliminating the blood: ongoing outbreaks of hepatitis B virus infection and the need for innovative glucose monitoring technologies. J Diabetes Sci Technol 2009; 3(2): Schaefer MK, Jhung M, Dahl M, et al: Infection control assessment of ambulatory surgical centers. JAMA 2010; 303(22): Kroon L: Overview of insulin delivery pen devices. J Am Pharm Assoc (2003) 2009; 49(5): Sonoki K, Yoshinari M, Iwase M, et al: Regurgitation of blood into insulin cartridges in the pen-like injectors. Diabetes Care 2001; 24(3): Le Floch JP, Herbreteau C, Lange F, Perlemuter L: Biologic material in needles and cartridges after insulin injection with a pen in diabetic patients. Diabetes Care 1998; 21(9): Centers for Disease Control and Prevention: Injection Safety. Infection Prevention during Blood Glucose Monitoring and Insulin Administration Available at accessed September 6, Kuehn BM: FDA warns against shared insulin pens. JAMA 2009; 301(15): U.S. Food and Drug Administration: Information for Healthcare Professionals: Risk of Transmission of Blood-borne Pathogens from Shared Use of Insulin Pens Available at PostmarketDrugSafetyInformationforPatientsandProviders/DrugSafety InformationforHeathcareProfessionals/ucm htm; accessed September 30,

9 18. Diabetes Health: Nassau Medical Center Extends Warning Over Insulin Pens Available at /nassau-medical-center-extends-warning-over-insulin-pens/; accessed September 6, Thompson ND, Perz JF, Moorman AC, Holmberg SD: Nonhospital health care-associated hepatitis B and C virus transmission: United States, Ann Intern Med 2009; 150(1): U.S. Public Health Service: Updated U.S. Public Health Service guidelines for the management of occupational exposures to HBV, HCV, and HIV and recommendations for postexposure prophylaxis. MMWR Recomm Rep 2001; 50(RR-11): Lessa F, Tak S, Devader SR, et al: Risk of infections associated with improperly reprocessed transrectal ultrasound-guided prostate biopsy equipment. Infect Control Hosp Epidemiol 2008; 29(4): Alter MJ, Kuhnert WL, Finelli L: Guidelines for laboratory testing and result reporting of antibody to hepatitis C virus. Centers for Disease Control and Prevention. MMWR Recomm Rep 2003; 52(RR-3): 1 13, 15; quiz CE Murphy DG, Willems B, Deschenes M, Hilzenrat N, Mousseau R, Sabbah S: Use of sequence analysis of the NS5B region for routine genotyping of hepatitis C virus with reference to C/E1 and 5 untranslated region sequences. J Clin Microbiol 2007; 45(4): Rubertone MV, Brundage JF: The Defense Medical Surveillance System and the Department of Defense serum repository: glimpses of the future of public health surveillance. Am J Public Health 2002; 92(12): Abramson JH: WINPEPI (PEPI-for-Windows): computer programs for epidemiologists. Epidemiol Perspect Innov 2004; 1(1): Dominitz JA, Boyko EJ, Koepsell TD, et al: Elevated prevalence of hepatitis C infection in users of United States veterans medical centers. Hepatology 2005; 41(1): Groom H, Dieperink E, Nelson DB, et al: Outcomes of a Hepatitis C screening program at a large urban VA medical center. J Clin Gastroenterol 2008; 42(1): Mallette C, Flynn MA, Promrat K: Outcome of screening for hepatitis C virus infection based on risk factors. Am J Gastroenterol 2008; 103(1): Armstrong GL, Wasley A, Simard EP, McQuillan GM, Kuhnert WL, Alter MJ: The prevalence of hepatitis C virus infection in the United States, 1999 through Ann Intern Med 2006; 144(10): Wasley A, Kruszon-Moran D, Kuhnert W, et al: The prevalence of hepatitis B virus infection in the United States in the era of vaccination. J Infect Dis 2010; 202(2): McQuillan GM, Kruszon-Moran D, Denniston MM, Hirsch R: Viral hepatitis. NCHS Data Brief 2010; 27: Weinbaum CM, Williams I, Mast EE, et al: Recommendations for identification and public health management of persons with chronic hepatitis B virus infection. MMWR Recomm Rep 2008; 57(RR-8): McCaughan GW, McGuinness PH, Bishop GA, et al: Clinical assessment and incidence of hepatitis C RNA in 50 consecutive RIBA-positive volunteer blood donors. Med J Aust 1992; 157(4): Dudzinski DM, Hebert PC, Foglia MB, Gallagher TH: The disclosure dilemma large-scale adverse events. N Engl J Med 2010; 363(10): Patel PR, Srinivasan A, Perz JF: Developing a broader approach to management of infection control breaches in health care settings. Am J Infect Control 2008; 36(10): William Beaumont Army Medical Center: Insulin Pen Advisory Available at ; accessed September 6,