Risk factors for explantation due to infection after sacral neuromodulation: a multicenter retrospective case-control study

Risk factors for explantation due to infection after sacral neuromodulation: a multicenter retrospective case-control study

Accepted Manuscript Risk factors for explantation due to infection after sacral neuromodulation: a multicenter retrospective case-control study Emily ...

NAN Sizes 0 Downloads 1 Views

Accepted Manuscript Risk factors for explantation due to infection after sacral neuromodulation: a multicenter retrospective case-control study Emily N.B. Myer, M.D, Andrey Petrikovets, M.D, Paul D. Slocum, M.D., Toy Gee Lee, M.D., Charelle M. Carter-Brooks, M.D., Nabila Noor, M.D., Daniela M. Carlos, M.D., Emily Wu, M.D., Kathryn van Eck, Ph.D., Tola B. Fashokun, M.D., Ladin YurteriKaplan, M.D., Chi Chiung Grace Chen, M.D, M.H.S PII:

S0002-9378(18)30282-5

DOI:

10.1016/j.ajog.2018.04.005

Reference:

YMOB 12129

To appear in:

American Journal of Obstetrics and Gynecology

Received Date: 15 January 2018 Revised Date:

23 March 2018

Accepted Date: 2 April 2018

Please cite this article as: Myer ENB, Petrikovets A, Slocum PD, Lee TG, Carter-Brooks CM, Noor N, Carlos DM, Wu E, van Eck K, Fashokun TB, Yurteri-Kaplan L, Chen CCG, Risk factors for explantation due to infection after sacral neuromodulation: a multicenter retrospective case-control study, American Journal of Obstetrics and Gynecology (2018), doi: 10.1016/j.ajog.2018.04.005. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT 1

Title:

2

Risk factors for explantation due to infection after sacral neuromodulation: a multicenter

3

retrospective case-control study

4

Authors:

5

Emily N.B. MYER, M.D.

RI PT

1

Johns Hopkins School of Medicine, Department of Gynecology and Obstetrics, Division of Female Pelvic Medicine and

7

Reconstructive Surgery

8

Andrey PETRIKOVETS, M.D. Case Western Reserve University School of Medicine,

SC

6

MetroHealth Medical Center, Department of Obstetrics and

M AN U

9 10

Gynecology, Division of Urogynecology and Pelvic

11

Reconstructive Surgery

12

Paul D. SLOCUM, M.D.

Vanderbilt University Medical Center, Department of Obstetrics and Gynecology, Division of Female Pelvic Medicine and Reconstructive Surgery

14 15

Toy Gee LEE, M.D.

19 20 21 22 23 24

EP

and Reconstructive Surgery

Charelle M. CARTER-BROOKS, M.D.

AC C

18

University of Rochester Medical Center, Department of Obstetrics and Gynecology, Division of Female Pelvic Medicine

16 17

TE D

13

Nabila NOOR, M.D.

Magee-Women's Hospital of University of

Pittsburgh Medical Center, Department of Obstetrics, Gynecology and Reproductive Sciences, Division of Urogynecology and Reconstructive Surgery Mount Auburn Hospital/Harvard Medical School, Department of Obstetrics and Gynecology

Daniela M. CARLOS, M.D. Montefiore Medical Center, Department of Obstetrics and

ACCEPTED MANUSCRIPT 2

25

Gynecology, Division of Female Pelvic Medicine and

26

Reconstructive Surgery Emily WU, M.D.

Scott and White Memorial Hospital, Department of Obstetrics and Gynecology

28 29

Kathryn VAN ECK, Ph.D.

RI PT

27

Johns Hopkins University School of Medicine Department of Psychiatry Division of Child and Adolescent Psychiatry and

31

Kennedy Krieger Institute Department of Psychiatry

32

SC

30

Tola B. FASHOKUN M.D.

Johns Hopkins School of Medicine, Department of Gynecology and Obstetrics, Division of Female Pelvic Medicine and

34

Reconstructive Surgery

35

Ladin YURTERI-KAPLAN, M.D.

M AN U

33

Columbia University Medical Center-New York

Presbyterian, Department of Obstetrics and Gynecology,

37

Division of Female Pelvic Medicine and Reconstructive Surgery

38

TE D

36

Chi Chiung Grace CHEN M.D, M.H.S Johns Hopkins School of Medicine, Department of Gynecology and Obstetrics, Division of Female Pelvic Medicine

39

and Reconstructive Surgery

EP

40 41

Corresponding Author:

43

Emily N.B. MYER

44 45

AC C

42

Johns Hopkins School of Medicine, Department of Gynecology and Obstetrics, Division of Female Pelvic Medicine and Reconstructive Surgery

46

Johns Hopkins Bayview Medical Center

47

4940 Eastern Avenue, 301 Building

ACCEPTED MANUSCRIPT 3

48

Department of Obstetrics and Gynecology

49

Baltimore MD 21224

50

(410)550-2787; fax (410)550-2786 The authors report no conflict of interest

52

Funding support provided by the AUGS-SGS Fellows Pelvic Research Network for the conduct

53

of research

54

Accepted for an oral poster at the 44th annual scientific meeting of the Society of Gynecologic

55

Surgeons, March 11-14, 2018, Orlando, Florida.

56

Word Count Abstract: 512

57

Word count main text: (3000 words):3209

M AN U

SC

RI PT

51

AC C

EP

TE D

58

ACCEPTED MANUSCRIPT 4

Abstract

60

Background: Sacral neuromodulation is an effective therapy for overactive bladder, urinary

61

retention, and fecal incontinence. Infection after sacral neurostimulation is costly and

62

burdensome. Determining optimal perioperative management strategies to reduce the risk of

63

infection is important to reduce this burden.

64

Objective: To identify risk factors associated with sacral neurostimulator infection requiring

65

explantation, to estimate the incidence of infection requiring explantation, and identify

66

associated microbial pathogens.

67

Study Design: This is a multicenter retrospective case-control study of sacral neuromodulation

68

procedures completed from January 1, 2004-December 31, 2014. We identified all sacral

69

neuromodulation implantable pulse generator implants (IPG) as well as explants due to infection

70

at eight participating institutions. Cases were patients who required IPG explantation for

71

infection during the review period. Cases were included if >18 years old, follow up data > 30

72

days after IPG implant, and if the implant was performed at the institution performing the

73

explant. Two controls were matched to each case. These controls were the patients who had an

74

IPG implanted by the same surgeon immediately preceding and immediately following the

75

identified case who met inclusion criteria. Controls were included if > 18 years old, no infection

76

after IPG implant, follow up data > 180 days after implant, and no explant for any reason <180

77

days from implant. Controls may have had an explant for reasons other than infection at >180

78

days after implant. Fishers exact (for categorical variables) and Student’s t-test (for continuous

79

variables) were used to test the strength of the association between infection and patient and

80

surgery characteristics. Significant variables were then considered in a multivariable logistic

81

regression model to determine risk factors independently associated with infection.

AC C

EP

TE D

M AN U

SC

RI PT

59

ACCEPTED MANUSCRIPT 5

82

Results: Over a ten-year period at 8 academic institutions, 1,930 sacral neuromodulator implants

84

were performed by 17 surgeons. Thirty-eight cases requiring device explant for infection and 72

85

corresponding controls were identified. The incidence of infection requiring explant was 1.97%.

86

Hematoma formation (13% cases, 0% controls; p=0.004) and pocket depth of >3cm (21% cases,

87

0% controls; p=0.031) were independently associated with an increased risk of infection

88

requiring explant. On multivariable regression analysis controlling for significant variables, both

89

hematoma formation (p=0.006) and pocket depth > 3cm (p= 0.020, odds ratio 3.26 [95%

90

confidence interval, 1.20, 8.89]) remained significantly associated with infection requiring

91

explant.

M AN U

SC

RI PT

83

92

Of the 38 cases requiring explant, 32 had cultures collected and 24 had positive cultures. All 5

94

cases with a hematoma had a positive culture (100%). Of the 4 cases with a pocket depth > 3cm,

95

2 had positive cultures, 1 had negative cultures and 1 had a missing culture result. The most

96

common organism identified was methicillin-resistant Staphylococcus aureus (38%).

EP

97

TE D

93

Conclusions: Infection after sacral neuromodulation requiring device explant is low. The most

99

common infectious pathogen identified was Methicillin-resistant Staphylococcus aureus.

AC C

98

100

Demographic and health characteristics did not predict risk of explant due to infection, however

101

having a postoperative hematoma or a deep pocket >3cm significantly increased the risk of

102

explant due to infection. These findings highlight the importance of meticulous hemostasis as

103

well as ensuring the pocket depth is <3cm at the time of device implant.

104

ACCEPTED MANUSCRIPT 6

105

Key Words: sacral neuromodulation, sacral nerve stimulation, explant, infection, overactive

106

bladder, risk factors

AC C

EP

TE D

M AN U

SC

RI PT

107

ACCEPTED MANUSCRIPT 7

108

Short title: Sacral neuromodulation infection risk factors requiring explantation

109

Condensation: Infection requiring explantation after sacral neuromodulator implant is low and

110

is associated with a deep implant pocket and postoperative hematoma formation.

112

Implications and Contributions:

113

Why was this study conducted?

114

*

115

neuromodulation.

RI PT

111

SC

To improve knowledge of risk factors for explantation due to infection after sacral

M AN U

116

What are the key findings?

118

*

The incidence of explantation due to infection is low.

119

*

Explantation is associated with postoperative hematoma formation and pocket depth

120

>3cm.

121

TE D

117

122

What does this study add to what is already known?

123

*

124

infection after sacral neuromodulation.

AC C

EP

Generalizable perioperative strategies that may reduce risk of explantation due to

ACCEPTED MANUSCRIPT 8

Introduction:

126

Sacral nerve stimulation (SNS) is a Food and Drug Administration (FDA) approved treatment

127

for patients with refractory overactive bladder, fecal incontinence, and non-obstructive urinary

128

retention. Since its introduction by Medtronic (Minneapolis, MN) in 1997, more than 150,000

129

InterStim® devices have been implanted and the use of SNS is increasing.1 Device implantation

130

is performed in two stages which has been previously reviewed.2 Briefly, patients first undergo a

131

testing phase whereby a lead is implanted into sacral foramen near the S3 nerve root. If there is a

132

50% or greater improvement in preoperative symptoms, an implantable pulse generator (IPG) is

133

then connected to a permanent lead and implanted under the skin of the lateral upper buttock.

M AN U

SC

RI PT

125

134

As of January 2017, the reported hospital charge for an IPG implant totaled $17,803.3 After a

136

testing phase 77-90% of patients move on to IPG implant.2,4,5 However, approximately 3-11% of

137

patients with an IPG implant have an infectious complication that may require explant.5-8 Device

138

explant is associated with significant morbidity including need for hospitalization for intravenous

139

antibiotics, surgical intervention to explant the device, higher cost and decreased quality of life.9

140

Additionally, the Centers for Medicare and Medicaid Services require mandatory reporting of

141

hospital-acquired surgical site infections and implantable device infections, which subsequently

142

impact hospital quality rating and reimbursement.10

EP

AC C

143

TE D

135

144

Infection of SNS requiring explant may be preventable. Patient risk factors associated with

145

surgical site infection that are not modifiable include age, obesity, and immunocompromised

146

status. However, modifiable factors that may decrease infection risk include antibiotic selection

147

and timing, wound closure technique, and length of testing phase.6 Most studies evaluating risk

148

factors for infection have been at a single institution with a specific practice for perioperative

ACCEPTED MANUSCRIPT 9

management to prevent infection limiting generalizability. Additionally, as infection is

150

uncommon, these studies have few cases limiting power to detect significant differences. The

151

objective of this study was to identify risk factors associated with SNS explantation due to

152

infection after placement of an IPG, to estimate the incidence of infection requiring explantation,

153

and identify the associated microbial pathogens. This was done as a multi-center study due to the

154

low a priori probability of explantation due to infection as well as to increase the generalizability

155

of our findings to inform perioperative management guidelines. We chose to include only cases

156

of infection requiring device explant as prior studies have reported on characteristics associated

157

with infection that may be treated without explant,11,12 and because explant is associated with

158

higher cost.

M AN U

SC

RI PT

149

159

Materials and Methods:

161

This retrospective case-control multicenter study, conducted through the Fellows’ Pelvic

162

Research Network (FPRN®), included eight clinical sites: Johns Hopkins University School of

163

Medicine, Baltimore, MD; Mount Auburn, Cambridge, MA; University of Rochester, Rochester,

164

NY; MetroHealth Medical Center /Case Western Reserve University School of Medicine,

165

Cleveland, OH; Vanderbilt University School of Medicine, Nashville, TN; Albert Einstein

166

College of Medicine, New York, NY; Scott & White, Temple, TX; and University of Pittsburgh,

167

Pittsburgh, PA. IRB approval was secured at each institution. Current procedural terminology

168

(CPT) code 64590 (insertion or replacement of peripheral or gastric neurostimulator pulse

169

generator or receiver, direct or inductive coupling) was used to identify all patients (male and

170

female) who had undergone implantation of a sacral neuromodulator device from January 1,

171

2004-December 31, 2014 at each institution. CPT codes 64585 (revision or removal of peripheral

AC C

EP

TE D

160

ACCEPTED MANUSCRIPT 10

neurostimulator electrode array) and 64595 (revision or removal of peripheral or gastric

173

neurostimulator pulse generator or receiver) were used to identify cases required IPG explant

174

during the review period. Record review was performed to identify cases where the IPG explant

175

was performed for infection. Infection was defined by the indication for explant being infection

176

in the preoperative and/or postoperative diagnosis of the explant procedure. All clinic,

177

perioperative nursing notes, operative notes, and the admission summary which includes record

178

of surgical skin prep, chlorhexidine wipes, antibiotics and other medications administered were

179

reviewed to assess diagnosis of infection and risk factors for infection.To confirm we captured

180

all patients who received chlorhexidine wipes in addition to record review, we also inquired each

181

sites institutional policy about use of chlorhexidine wipes before surgery and 2 institutions

182

routinely used and therefore the cases and controls from these institutions were considered as

183

using preoperative chlorhexidine wash. Subjects may have had an IPG implant either as part of a

184

staged implant (permanent lead followed by IPG implant) or a full implant (temporary lead

185

followed by permanent lead with IPG implant). Cases were included if >18 years old, follow up

186

data > 30 days after IPG implant, and implant was performed at the institution performing the

187

explant. Two controls were matched to each case. These controls were the patients who had an

188

IPG implanted by the same surgeon immediately preceding and immediately following the

189

previously identified case. Controls were included if > 18 years old, no infection after IPG

190

implant as verified by record review of postoperative clinic notes showing no signs or symptoms

191

of infection and no diagnosis of infection by the physician, follow up data > 180 days after

192

implant, and no explant for any reason <180 days from implant. Controls may have had an

193

explant for reasons other than infection at >180 days after IPG implant. To determine the

194

incidence of explantation due to infection, we identified all IPG implantation procedures and

AC C

EP

TE D

M AN U

SC

RI PT

172

ACCEPTED MANUSCRIPT 11

195

explantation procedures due to infection at each of the eight participating institutions over the

196

10-year review period.

197

Patient demographic information including age, gender, body mass index (BMI), indication for

199

SNS, Charlson medical comorbidity index (CCMI), composite immunocompromising risk

200

factors (diabetes, steroids, tobacco use and acquired immune deficiency syndrome (AIDS)),

201

methicillin-resistant Staphyloccocus aureus (MRSA) carrier status, and use of anticoagulants

202

was recorded. Recent hemoglobin A1c was recorded for diabetic patients. Institutional and

203

surgeon data including the presence of learners (fellows and residents), number of implants each

204

surgeon performed within the review period, and surgeon specialty was recorded. Intraoperative

205

data from implant procedures including number of leads and IPGs placed, days between lead and

206

IPG implant, intraoperative intravenous antibiotic prophylaxis, use of irrigation antibiotic

207

solution, postoperative antibiotic prophylaxis, IPG pocket depth recorded as incision depth from

208

skin, pocket closure method and postoperative complications including hematoma formation as

209

diagnosed clinically after IPG implant were identified. Finally, we identified data related to the

210

infection and explantation procedure including cultures, antibiotic use, hospitalization, timing of

211

explant after implantation or antibiotic use, and whether or not re-implantation was performed.

SC

M AN U

TE D

EP

AC C

212

RI PT

198

213

Although there are no studies that specifically addresses explantation due to infection, Bruseke et

214

al. found that the risk factor most highly associated with overall infection was

215

immunocompromised status. The overall prevalence of immunocompromised status in their

216

population was 25% and the overall prevalence of infection was 3%. However, the authors did

217

not report the specific difference in the rates of infection between individuals with and without

ACCEPTED MANUSCRIPT 12

immunosuppression. Therefore, for our power analysis assuming a type I error of 5%, power

219

level of 0.8, and 1:2 case-control ratio, we used the immunocompromised status as the exposure

220

of interest and the aforementioned prevalence estimates to calculate the sample sizes required to

221

detect a 2-fold increase, a 3-fold increase, and a 4-fold increase in the odds of explantation in the

222

immunocompromised individuals and found sample sizes of 347 (119 cases, 238 controls), 297

223

(99 cases, 198 controls), and 81 (27 cases, 54 controls), respectively. Although we did not know

224

the exact rate of explantation, we hypothesize that it is an uncommon event and decided to use

225

the sample size estimation of 27 cases and 54 controls.

M AN U

226

SC

RI PT

218

Fishers exact and Student's t-test tests were used to test the strength of the association between

228

explant due to infection infection and subject and surgery characteristics for categorical and

229

continuous variables, respectively. Significant variables were then considered in a multivariable

230

logistic regression model to determine risk factors independently associated with infection. We

231

chose to include only significant variables in the multivariable model to improve power due to

232

the low incidence of infection requiring explantation. Statistical analysis was completed with

233

STATA (StataCorp 2015, version 14, College Station, TX). Statistical significance was defined

234

at the 5% significance level.

EP

AC C

235

TE D

227

236

Results

237

A total of 1,930 IPG implant procedures and 38 cases of infection requiring IPG explant were

238

performed by 17 surgeons across a 10-year review period at eight FPRN® sites. The incidence of

239

infection requiring IPG explant was 1.97%. From these implanted patients, 72 matched controls

240

to the 38 cases were identified resulting in a total sample size of 110. Four cases had

ACCEPTED MANUSCRIPT 13

overlapping controls serving both as the preceding and following implant or cases with no other

242

nearest implant to serve as an alternative control. The most common reported signs of infection

243

were pain (87%), erythema (79%), induration (63%) and purulent drainage (50%); fever (2%)

244

and sepsis (1%) were uncommon. Of the 38 cases requiring explant, 32 had cultures collected

245

and 24 had positive cultures. All cases with a hematoma had a positive culture (100%). Of the 4

246

cases with a pocket depth > 3cm, 2 cases cultures were positive, 1 case was negative and 1 case

247

was missing the culture result. The most common organisms identified were MRSA (38%) and

248

methicillin-sensitive staphylococcus aureus (33%) followed by Pseudomonas (10%),

249

Staphylococcus epidermidis (10%), Proteus (5%), and Corynebacterium (5%). Treatment for

250

infection prior to explanation included outpatient antibiotics (89%) and admission with

251

intravenous antibiotics (18%). The median time between implant of IPG and explant for

252

infection was 53 days (interquartile range (IQR), 21,150). For cases who had an antibiotic trial

253

prior to explant (n=33, 89%), the time to explant was on average 14 days after the trial of

254

antibiotics (IQR, 5, 30). Re-implantation was performed in 42% of cases a median of 93 days

255

after explant (IQR, 60, 365). Seven controls (9.7%) had a device explantation performed at >180

256

days after implant for indications of ineffective (n=5), pain (n=1), patient desired (n=1); none

257

were considered to have an infection prior to or at the time of device removal.

SC

M AN U

TE D

EP

AC C

258

RI PT

241

259

All surgical procedures were performed at teaching institutions with residents and/or fellows

260

participating in surgery. Subject demographics are summarized in table 1. There were no

261

significant differences in gender, race, BMI, age CCMI, or indication for implantation between

262

cases and controls (table 1). Composite immunocompromised status was not significantly

263

associated with explantation due to infection (59% cases, 53% controls, p=0.547). Of the 8

ACCEPTED MANUSCRIPT 14

cases (21%) and 18 controls (25%) with diabetes, there was no difference in median A1c (cases

265

7.2, IQR, 6.2, 8.7; controls 7, IQR, 5.7, 7.1; p=0.675). MRSA carrier status was not associated

266

with infection requiring explant (6% cases, 0% controls, p=0.118. Median surgeon volume

267

during the 10 year review period was 235 implants (IQR, 71, 278). Surgeon specialty was not

268

associated with infection requiring explant (p=0.955) (table 1). Temporary lead testing phases

269

were performed in 3 cases and 6 controls and was not associated with infection requiring explant

270

(p=1.00).

SC

RI PT

264

271

At stage I, there was no association between skin prep solution, intraoperative, postoperative or

273

irrigation antibiotic, number of leads, side of lead versus pocket, or infection of leads before IPG

274

implant and risk of explant due to infection (table 2). Two cases were diagnosed with a lead

275

infection after stage I which were treated with lead explant and outpatient antibiotics. These

276

cases went on to have a full implant (lead + IPG) after the infection was considered resolved.

277

Both cases ultimately required explant of the IPG due to infection.

TE D

278

M AN U

272

The median time between stage I and IPG implant was 14 days for both cases and controls (IQR,

280

11, 14 and 12, 14, respectively) (table 3). At the time of IPG implant, infection requiring explant

281

was not associated with removal and/or placement of leads, side of lead in relation to pocket,

282

skin prep solution, intraoperative antibiotic, use of irrigation antibiotics, or post procedural oral

283

antibiotic prophylaxis (table 3). Few patients used chlorhexidine wipes the night before surgery

284

(8% cases, 10% controls). Suture material to close the pocket was not associated with infection

285

(p=0.874). Pocket depth >3cm was significantly associated with infection requiring explantation

286

(21% cases, 0% controls p=0.031); though IPG pocket depth was only recorded in 50% of cases

AC C

EP

279

ACCEPTED MANUSCRIPT 15

and 44% of controls. There was no association between BMI and pocket depth (p=0.728) with

288

obesity being present in 75% of subjects with pocket depth >3cm, 50% of subjects with pocket

289

depth 2-3cm and 58% of subjects with pocket depth 1-2cm. There was also a significant

290

association between infection requiring explant and postprocedural hematoma (p=0.004) which

291

was identified in 5 cases (13%) and no controls, one of which required surgical drainage. No

292

patients who developed a hematoma used chronic anticoagulation. One patient with a deep

293

pocket also developed a postprocedureal heatoma, 3 were in pockets 1-2cm and 1 had an

294

unrecorded pocket depth. On multivariable regression analysis controlling for hematoma and

295

pocket depth, both hematoma formation (p=0.006) and pocket depth > 3cm (p= 0.020, odds ratio

296

3.26 [95% confidence interval, 1.20, 8.89]) remained significantly associated with infection

297

requiring IPG explant.

AC C

EP

TE D

M AN U

SC

RI PT

287

ACCEPTED MANUSCRIPT 16

Comment:

299

We found a low rate of infection requiring explant of SNS of 1.97% with MRSA being the most

300

common organism. Interestingly, a pocket depth > 3cm and hematoma formation were

301

independently associated with infection requiring explant, which has not been reported in other

302

studies. We found no statistically significant difference in risk of infection requiring explant

303

based on other health or surgery characteristics.

SC

304

RI PT

298

Our low rate of infection and infecting organisms are consistent with other studies.11,13,14 We

306

would expect our rate to be lower than other studies as we included only cases of infection

307

requiring explant and excluded infection treated successfully with antibiotics.11,12 BMI or

308

immunocompromised status as was reported by Brueseke et al8 was not significantly associated

309

with a risk of explanatation due to infection as both the explanation group (cases) and the non-

310

explantation group (controls) had similarly high rates of these patients.

TE D

311

M AN U

305

We found no difference in risk of infection requiring explant based on length of testing phase;

313

this may be because most patients had a testing period of <14 days and >14 day testing period

314

was a risk factor for infection in other studies.15 Though infection after the testing phase before

315

IPG implant was not significantly associated with infection requiring explant, this is likely

316

because there were too few cases to detect a difference and the only two cases with such an

317

infection both required IPG explant due to infection. Too few patients had a known MRSA

318

carrier status to detect any association between MRSA status and infection, though MRSA was

319

the most common organism identified.

320

AC C

EP

312

ACCEPTED MANUSCRIPT 17

Both pocket depth > 3cm and hematoma formation were independent risk factors for infection

322

requiring explant. It was unexpected that a deeper pocket would be associated with an increased

323

risk of infection as theoretically a more a superficial pocket would be at greater risk of device

324

exposure and subsequent infection. It is possible that a deeper pocket leads to a larger potential

325

space for fluid accumulation resulting in subsequent seroma formation, infection, and wound

326

separation.16 There was not association between pocket depth and BMI, though the low number

327

of cases with a deep pocket limits the power to detect a difference. Additionally, pocket depth

328

was only recorded in 50% of cases and controls. It is not surprising that wound hematomas were

329

associated with infection requiring explant as in addition to blood being an excellent

330

environment for bacterial growth, the pressure of the hematoma against the wound can reduces

331

tissue oxygenation and delay wound healing.17 These findings support the importance of

332

meticulous hemostasis at the time of IPG pocket formation as well as making the IPG pocket

333

<3cm.

SC

M AN U

TE D

334

RI PT

321

In regards to antibiotic use, the Infectious Disease Society of America (ISDA) recommends

336

prophylaxis with cefazolin ± aminoglycoside, cefazolin ± aztreonam, or ampicillin–sulbactam;

337

alternative prophylaxis for allergies includes clindamycin ± aminoglycoside or aztreonam,

338

vancomycin ± aminoglycoside or aztreonam.19 We found no difference in risk of infection based

339

on choice ofantibiotic at implant with >50% of cases and controls having cefazolin-only

340

antibiotic prophylaxis. This is in contrast to a study by Haraway et al who found a 7-time higher

341

odds of infection with cefazolin only prophylaxis as compared to vancomycin prophylaxis in a

342

review of 136 implants and 8 infection cases.14 There is insufficient data to support the use of

343

postoperative oral antibiotic prophylaxis, nor is it recommended by the device manufacturer, yet

AC C

EP

335

ACCEPTED MANUSCRIPT 18

80% of high-volume providers use postoperative antibiotics to prevent surgical site infection.20

345

With high rates of postoperative antibiotic prophylaxis (68% cases, 60% controls) and antibiotic

346

irrigation solution ( 68% cases, 71% controls) at the time of implant, we found no significant

347

association between antibiotic prophylaxis and infection requiring explant. Given the high use in

348

our study, in an era of multidrug resistant bacteria, further studies are necessary to validate the

349

benefit of these prophylaxis measures.

RI PT

344

SC

350

The primary limitations of our study include the uncommon event of explant due to infection

352

limiting power to detect significant differences for the variables of interest. We did not record

353

use of intraoperative anticoagulant prophylaxis as given the short procedure time it seemed

354

unlikely to be used at most institutions, but this would be a consideration for future evaluation

355

given the association of hematoma and infection . Additionally, it is possible there were missed

356

cases of asymptomatic postoperative hematomas after IPG implant in control subjects. An

357

additional limitation of our study was that only 50% of patients had pocket depth recoreded in

358

the operative note which limits the generalizability of our findings. We did not assess operating

359

time for stage I or stage II procedures and cannot assess the risk of length of surgery on infection

360

rate in this study which was noted to be a risk factor for infection by Guralnick et al.13 Due to

361

low use in our study, we are unable to determine if use of an gentamicin-impregnated collagen

362

sheet or chlorehexadine wash prior to surgery would reduce the risk of infection as has been seen

363

in single-institution retrospective chart reviews.11,21 We also cannot comment on the risk factors

364

for infection treated with antibiotics as we did not include these patients in our review.

365

AC C

EP

TE D

M AN U

351

ACCEPTED MANUSCRIPT 19

The major strengths of our study are the multicenter design to increase management, surgeon and

367

patient diversity and generalizability of the results. We had a large number of case subjects with

368

sufficient immunocompromising risk factors to detect a difference, if present. Additionally, we

369

evaluated the efficacy of various antibiotic prophylaxis regimens (intravenous, irrigation, oral

370

postoperative) at both stages I & II to reduce the risk of infection.

371

RI PT

366

Other specialties with implantable devices, specifically the American Heart Association for

373

implantable cardiac electronic devices, have developed guidelines for the diagnosis, management

374

and prevention of implantable device infection.22 We propose that a similar guideline be

375

developed for sacral neuromodulation. Based on our findings, we would suggest the following:

376

1. Use of either betadine or chlorhexidine for skin preparation, 2. Administer intraoperative

377

antibiotic prophylaxis in line with ISDA recommendations 3. Consider developing a pocket

378

depth of less than 3 cm at the time of IPG placement 4. Maintain meticulous hemostasis to avoid

379

subsequent hematoma formation. We further recommend the preoperative use of chlorhexaine

380

wipes based on the review by Brueseke et al showing a 5.7% reduction in infection (7.4% to

381

1.7%, p=0.002). Based on the existing literature and our study, there is not sufficient evidence to

382

support the routine use of intraoperative irrigation with antibiotic solution or postoperative

383

antibiotic prophylaxis.

M AN U

TE D

EP

AC C

384

SC

372

ACCEPTED MANUSCRIPT 20

385

References

387

1. Laudano MA, Seklehner S, Sandhu J, et al. Disparities in the use of sacral neuromodulation

388

among medicare beneficiaries. J Urol. 2015;194(2):449-453.

389

2. Noblett K, Cadish L.

390

Sacral nerve stimulation for the treatment of refractory voiding and bowel dysfunction Am J

391

Obstet Gynecol. 2014;210:99-106.

392

3. Medtronic. Commonly billed codes sacral neuromodulaon for urinary control and bowel

393

control commonly billed code. http://www.medtronic.com/content/dam/medtronic-

394

com/professional/documents/uc201002977m-snm-reimbursement-coding-2017.pdf. Updated

395

2017.

396

4. Siegel S, Noblett K, Mangel J, et al. Five-year follow up results of a prospective,

397

multicenterstudy of patients with overactive bladder treated with sacral neuromodulation. J Urol.

398

2018;199:229-236.

399

5. Amundsen C, Komesu Y, Chermansky C, et al. Two-year outcomes of sacral neuromodulation

400

versus OnabotulinumtoxinA for refractory urgency urinary incontinence: A randomized trial.

401

Eur Urol. 2018;Epub.

402

6. Lee C, Pizarro-Berdichevsky J, Clifton M, Vasavada S. Sacral neuromodulation implant

403

infection: Risk factors and prevention. Curr Urol Rep. 2017;18(2):16:1-6.

AC C

EP

TE D

M AN U

SC

RI PT

386

ACCEPTED MANUSCRIPT 21

7. Noblett K, Benson K, Kreder K. Detailed analysis of adverse events and surgical interventions

405

in a large prospective trial of sacral neuromodulation therapy for overactive bladder patients .

406

Neurourol Urodyn. 2017;36:1136-1139.

407

8. Hull T, Giese C, Wexner S, et al. Long-term disability of sacral nerve stimulation therapy for

408

chronic fecal incontinence. Dis Colon Rectum. 2013;56:234-245.

409

9. Das A, Carlson A, Hull M.

410

Improvement in depression and health-related quality of life after sacral nerve stimulation

411

therapy for treatment of voiding dysfunction Urology. 2004;64:62-68.

412

10. Centers for Medicaid and Medicare Services. Hospital-acquired conditions. Centers for

413

Medicaid and Medicare Services. Web site. https://www.cms.gov/Medicare/Medicare-Fee-for-

414

Service-Payment/HospitalAcqCond/icd10_hacs.html. Updated 08/2017.

415

11. Brueseke T, Livingston B, Warda H, Osann K, Noblett K.

416

Risk factors for surgical site infection in patients undergoing sacral nerve modulation therapy

417

Female Pelvic Med Reconstr Surg. 2015;21(4):198-204.

418

12. Wexner S, Hull T, Edden Y, et al.

419

Infection rates in a large investigational trial of sacral nerve stimulation for fecal incontinence J

420

Gastrointest Surg. 2010;14:1081-1089.

421

13. Guralnick M, Benouni S, O'Connor R, Edmiston C.

422

Characteristics of infections in patients undergoing staged implantation for sacral nerve

423

stimulation Urology. 2007;69:1073-1076.

AC C

EP

TE D

M AN U

SC

RI PT

404

ACCEPTED MANUSCRIPT 22

14. Haraway A, Clemens J, He C, Stroup C, Atiemo H, Cameron A.

425

Differences in sacral neuromodulation device infection rates based on preoperative antibiotic

426

selection Int Urogynecol J. 2013;24:2081-2085.

427

15. Lai HH, Grewal S. Bacterial colonization rate of InterStim and infection outcome with staged

428

testing. Urology. 2013;82(6):1255-1260.

429

16. Schweinberger M, Roukis T. Wound complications. Clin Podiatr Med Surg. 2009;26:1-10.

430

17. Wilson J, Clark J. Obesity: Impediment to wound healing. Crit Care Nurs Q. 2003;262:119-

431

132.

432

18. Chughtai B, Sedrakyan A, Isaacs A, Lee R, Te A, Kaplan S.

433

Long term safety of sacral nerve modulation in medicare beneficiaries Neurourol Urodyn.

434

2015;34:659-663.

435

19. Bratzler DW, Dellinger EP, Olsen KM, et al. Clinical practice guidelines for antimicrobial

436

prophylaxis in surgery. Surg Infect (Larchmt). 2013;14(1):73-156.

437

20. Lee E, Lucioni A, Lee U, Kobashi K. National practice patterns of infection prophylaxis for

438

sacral neuromodulation device: A survey of high volume providers. Urol Pract. 2015;2:38-43.

439

21. Simpson JA, Peacock J, Maxwell-Armstrong C. Use of a gentamicin-impregnated collagen

440

sheet (collatamp((R)) ) following implantation of a sacral nerve stimulator for faecal

441

incontinence. Colorectal Dis. 2012;14(4):e200-2.

AC C

EP

TE D

M AN U

SC

RI PT

424

ACCEPTED MANUSCRIPT 23

442

22. Harrison JL, Prendergast BD, Sandoe JA. Guidelines for the diagnosis, management and

443

prevention of implantable cardiac electronic device infection. Heart. 2015;101(4):250-252.

RI PT

444

AC C

EP

TE D

M AN U

SC

445

ACCEPTED MANUSCRIPT 24

Table 1: Demographic and surgery characteristics of study population Demographics

Cases

Controls

N=38

N=72

p-value

0.414

RI PT

Gender 3(8)

3(4)

Female

35 (92)

70(96)

55.5 (47-61)

60 (43-71)

Age at implant Race

33(92)

Non-white

3(8)

BMI

19(26)

25-<30 kg/m2

5(13)

13(18)

25(66)

40(56)

TE D

8(21)

Diagnosis n (%)

0.640

0.940

29(76)

56(78)

Non-obstructive urinary retention

4(11)

8(11)

Fecal incontinence

5(13)

8(11)

22(59)

38(53)

EP

Overactive bladder (urgency,

0.248

12(18)

<25 kg/m2

>30 kg/m2

0.292

56(82)

M AN U

White

SC

Male

frequency, urge incontinence)

AC C

446

Comorbidities

Composite immunocompromised status (any diabetes, steroid use, AIDS, tobacco)

0.547

ACCEPTED MANUSCRIPT 25

7.2(6.15-8.7)

7(5.7-7.1)

0.675

CCMI, total score, median (IQR)

2(1-4)

2.5(0-4)

0.856

MRSA carrier ±

2(6)

0(0)

0.118

Anticoagulation use

1(3)

5(7)

0.660

Surgery Characteristics Surgeon training

0.955

23(61)

Urology (general)

11(29)

38(100)

First Procedures performed

19(26) 8(11)

M AN U

Teaching institution

45(63)

SC

FPRMS board certified

Other (Colorectal, General Ob/Gyn) 4(11)

72(100)

3(8)

6(8)

Staged implant, stage I followed by

66(92)

35(92)

TE D

Peripheral nerve evaluation

stage II 447

RI PT

Hemoglobin A1C (in diabetics)

1.00 1.00

CCMI Charleston Comorbidity Index; AIDS acquired immunodeficiency syndrome; BMI body

449

mass index; MRSA methicillin-resistant Staphylococcus aureus; Data are recorded as n (%) or

450

median (interquartile range). N=110

451

± reported positive in only 2 cases with documentation of carrier status in medical record, other

452

subjects had no recorded testing performed

453

AC C

EP

448

ACCEPTED MANUSCRIPT 26

454

Table 2: Stage I sacral neuromodulation perioperative management characteristics of cases

455

and controls

Home chlorhexidine wash before surgery

Cases

Controls

N=35

N=69

2(10)

6(12)

1.00

25(71)

Betadine-Iodine

9(26)

Other

1(3)

Intraoperative antibiotic, stage I Cephalosporins Aminoglycosides

Other/None

TE D

Penicillins

48(74)

16(25)

1(2)

M AN U

Chlorhexadine

Vancomycins

1.00

SC

Skin prep solution

p-value

RI PT

Characteristic

EP

Antibiotic irrigation solution used (5%

0.727

19(58)

37(57)

1(3)

4(6)

1(3)

0(0)

11(33)

20(30)

1(3)

4(6)

23(66)

41(60)

0.591

22 (63)

41 (59)

0.735

AC C

gentamicin, bacitracin & other)

Post-procedure prophylactic antibiotic used Number of days treated with post-procedure

1.00

prophylactic antibiotics, explant, stage I <7 days

3(20)

5(18)

>7 days

12(80)

23(82)

Number of leads placed at stage I, n(%)

1.00

ACCEPTED MANUSCRIPT 27

Unilateral

23(65)

44(65)

Bilateral

12(34)

24(35) 0.093

placed)

RI PT

Site of pocket in relation to lead (if 1 lead

17(68)

26(46)

Opposite side

8(32)

30(54)

Infection suspected after stage I

2(5)

0(0)

0.107

Stage I infection treated with explant of leads

2(100)

--

--

Stage I infection treated with outpatient

2(100)

--

--

M AN U

antibiotics

SC

Same side

Data are recorded as n (%) or median (interquartile range). N=104 as 3 cases and 3 controls

457

underwent successful temporary lead testing phase and did not have an independent stage I

458

implant.

AC C

EP

TE D

456

ACCEPTED MANUSCRIPT 28

459

Table 3: Internal pulse generator (IPG) implant perioperative management characteristics

460

and explant characteristics of cases and controls Controls

N=38

N=72

14(11-14)

14(12-14)

0.710

15(15-15)

--

--

5(9.8)

1.00

implant, median Days between stage I and IPG implant if infection of stage I

Skin prep solution Chlorhexadine Betadine-Iodine

Penicillins

AC C

Vancomycins

EP

Aminoglycosides

TE D

Intraoperative antibiotic Cephalosporins

2(8)

M AN U

Home chlorhexidine wipes before surgery

Other/None

Antibiotic irrigation solution used (5%

p-value

RI PT

Number of days between stage I and IPG

Cases

SC

IPG implant

23(66)

49(71)

12(34)

20(29)

0.655

0.250

16(47)

35(55)

4(12)

9(10)

1(3)

0(0)

10(290)

22(34)

3(9)

1(2)

26(68)

51(71)

0.829

26(68)

43(60)

0.412

gentamicin, bacitracin, other) Post-procedure prophylactic antibiotic used Number of days treated with post-procedure prophylactic antibiotics

1.00

ACCEPTED MANUSCRIPT 29

<7 days

3(20)

5(18)

>7 days

12(180)

23(82)

Number of leads removed at IPG implantˆ

0.312 30(79)

52(72)

Unilateral

8(21)

15(21)

Bilateral

0(0)

5(7)

24(63)

Unilateral

7(18)

Bilateral

7(18)

Depth of pocket ± 1-2cm

TE D

2-3 cm >3 cm

14(19) 4(6)

9(47)

22(69)

6(32)

10(31)

4(21)

0(0)

AC C

Same side

EP

Side of pocket relative to lead, if unilateral lead

Opposite side

Estimated blood loss

0.031

0.430

15(68)

27(56)

7(32)

21(44)

4(0-10)

3(1-5)

Hematoma formation

0.439 0.004

Yes

5(13)

0(0)

No

33(87)

72(100)

Hematoma treatment

0.106

54(75)

M AN U

None

SC

Number of leads placed at IPG implant§

RI PT

None

--

ACCEPTED MANUSCRIPT 30

Expectant management

4(80)

--

Surgical drainage

1(20)

--

Data are recorded as n (%) or median (interquartile range)

462

ˆ indications for lead removal: lead revision, ineffective lead (if 2 leads placed at testing phase)

463

§

464

leads placed), lead revision, replacement of ineffective lead, full implant in 2 cases with infection

465

after stage I who subsequently had full implant of lead(s) with IPG after resolution of infection.

466

±

RI PT

461

data available for 50% of cases and 44% of controls

AC C

EP

TE D

M AN U

467

SC

indications for lead placement: need for permanent lead after successful testing phase (1 or 2