Routine Workup of Postoperative Pyrexia Following Total Joint Arthroplasty Is Only Necessary in Select Circumstances

Accepted Manuscript Routine Workup of Postoperative Pyrexia following Total Joint Arthroplasty is Only Necessary in Select Circumstances Je-Hyun Yoo, MD, PhD, Camilo Restrepo, MD, Antonia F. Chen, MD, MBA, Javad Parvizi, MD, FRCS PII:

S0883-5403(16)30668-4

DOI:

10.1016/j.arth.2016.09.025

Reference:

YARTH 55417

To appear in:

The Journal of Arthroplasty

Received Date: 9 August 2016 Revised Date:

19 September 2016

Accepted Date: 20 September 2016

Please cite this article as: Yoo J-H, Restrepo C, Chen AF, Parvizi J, Routine Workup of Postoperative Pyrexia following Total Joint Arthroplasty is Only Necessary in Select Circumstances, The Journal of Arthroplasty (2016), doi: 10.1016/j.arth.2016.09.025. 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

Je-Hyun Yoo MD, PhD Camilo Restrepo MD Antonia F. Chen MD, MBA Javad Parvizi MD, FRCS

RI PT

Routine Workup of Patients with Postoperative Pyrexia following Total Joint Arthroplasty is Not Necessary

EP

TE D

M AN U

SC

The Rothman Institute at Thomas Jefferson University, Philadelphia, PA

AC C

Corresponding Author:

Javad Parvizi MD, FRCS The Rothman Institute at Thomas Jefferson University 125 South 9th St. Ste 1000 Philadelphia, PA 19107 P: (267) 339-3572 F: (215) 503-5651 [email protected]

ACCEPTED MANUSCRIPT 1

Routine Workup of Postoperative Pyrexia following Total Joint Arthroplasty is Only

2

Necessary in Select Circumstances

3

Running Title: Postoperative Pyrexia following TJA

RI PT

4

AC C

EP

TE D

M AN U

SC

5

1

ACCEPTED MANUSCRIPT Abstract

7

Background: It is unclear when routine workup of postoperative pyrexia (POP) following

8

total joint arthroplasty (TJA) should be performed.

9

Methods: A retrospective electronic database search was conducted on 25,558 consecutive

10

patients undergoing primary or revision TJA between June 2001 and June 2013. We identified

11

patient demographics, procedure type, characteristics of feverish patients, and febrile

12

complications. The estimated costs for chest x-ray (CXR), urinalysis (UA), urine culture

13

(UC), and blood culture (BC) were investigated.

14

Results: POP occurred in 46% of TJAs. A total of 11,589 separate workups were performed in

15

90.5% of POP patients, of which 2.4% were positive. UA, UC, BC, and CXR were positive in

16

38.7%, 9.5%, 7.0%, and 0.2%, respectively. Febrile complications occurred in 4.5% and the

17

infectious complications rate was 2.0%. The positive rate of fever workups was significantly

18

higher in patients with the first POP occurring after postoperative day (POD) 3, POP>102°F,

19

multiple fever spikes, and patients undergoing revision TJA. Multivariate logistic regression

20

revealed that the time of first POP, the maximum temperature, multiple fever spikes, and

21

revision TJA were independent predictors of febrile complications. The estimated cost for

22

11,319 negative workups in patients with POP was $4,636,976.80, with CXR costing

23

$4,613,182.00.

24

Conclusion: Selective workup of POP following TJA should be performed in patients with

25

higher temperatures, fever occurring after POD 3, those with multiple fever spikes, and those

26

undergoing revision TJA. CXR with an extremely low positive rate should not routinely be

27

ordered.

AC C

EP

TE D

M AN U

SC

RI PT

6

28 29 30

Keywords: total joint arthroplasty; postoperative pyrexia; routine workup; febrile complication 2

ACCEPTED MANUSCRIPT 31

Introduction: Postoperative pyrexia (POP) is common following total joint arthroplasty (TJA) [1-

33

3]. To detect potentially serious complications such as periprosthetic joint infection (PJI),

34

routine workup to determine the cause of POP for patients undergoing TJA is performed. A

35

previous study suggested that pyrexia on postoperative days (POD) 1 and 2 is often non-

36

infectious and most likely benign, whereas the presence of fever on POD 3, 4, or 5 may have

37

an infectious etiology [4]. POP may be due to septic and aseptic causes such as PJI, urinary

38

tract infection (UTI), surgical site infection (SSI), pneumonia, atelectasis, pulmonary

39

embolism (PE), deep venous thrombosis (DVT), and administration of some medications.

40

Timely workup for POP is essential for early diagnosis of infection after TJA, as well as

41

providing appropriate management of these complications [3].

M AN U

SC

RI PT

32

Routine evaluation of all patients with POP may, however, be ineffective and

43

wasteful [5-9]. Several authors have reported that POP develops after TJA due to a normal

44

inflammatory reaction from tissue injury during surgery and does not necessitate a diagnostic

45

workup [10-12]. However, POP has been clinically linked to delays in hospital discharge and

46

increased healthcare costs. Based on the available literature, the cause of POP after TJA

47

remains unclear. Furthermore, it is not known how often these fever workups are positive and

48

if the tests should routinely be performed.

EP

AC C

49

TE D

42

Using a large patient cohort, this study was designed to: a) Investigate the incidence

50

and characteristics of POP in the early postoperative period following TJA; b) Evaluate the

51

positive results, timing, and cost-effectiveness of specific diagnostic tests for POP; and c)

52

Suggest a modified, evidence-based algorithm for management of POP after TJA.

53 54 55

Materials and Methods A retrospective study was performed on 25,558 patients who underwent primary or 3

ACCEPTED MANUSCRIPT revision TJA at our institution between June 2001 and June 2013. Approval from the

57

institutional review board was obtained prior to initiation of the study. All patients received

58

antibiotic prophylaxis within one hour of skin incision using cefazolin 1-2g, or vancomycin

59

1-1.5g, if allergic to cefazolin. Antibiotics were then continued for 24 hours postoperatively

60

using either cefazolin every 8 hours or vancomycin every 12 hours. We obtained body

61

temperatures by placing a thermometer under the tongue of patients, per hospital routine. For

62

this study, we defined POP as any body temperature of 100.4°F (38°C) or higher in the early

63

postoperative period occurring between POD 0 and 7 [5,6,8]. POP in each patient was

64

characterized by the maximum temperature (Tmax), the day of the first fever, and frequency

65

of fevers, stratified as either single or multiple fever spikes. In addition, the occurrence of

66

POP was investigated according to the type of arthroplasty performed.

M AN U

SC

RI PT

56

According to our algorithm for treatment of POP after TJA (Figure 1), serial

68

temperatures >100.4°F and ≤102°F within 24 hours after surgery were treated with

69

acetaminophen. Patients with temperatures >102°F or those who had fever over 24 hours

70

after surgery were worked up with our institutional protocol that included chest x-ray (CXR)

71

to rule out atelectasis or pneumonia, followed by urinalysis (UA), urine culture (UC), and

72

blood culture (BC). In patients with persistent POP and negative CXR, UC, and BC, other

73

interventions may have been performed that included joint aspiration, Doppler ultrasound, or

74

chest computed tomography (CT) to exclude SSI, PJI, DVT, or PE. We defined septic or

75

aseptic febrile complications as pneumonia, UTI, PJI, sepsis, SSI, DVT, and PE related to the

76

index surgery. PJI was diagnosed according to the Musculoskeletal Infection Society criteria

77

[13].

78 79 80

AC C

EP

TE D

67

The diagnostic workup for POP was deemed to be positive using the following definitions [14-16]: 1) UA was deemed to be positive if the strips for leukocyte esterase or nitrite on the urine 4

ACCEPTED MANUSCRIPT 81 82

83

dip stick were positive or 3 to 5 white blood cells per high-power field were seen. 2) UC was deemed to be positive if ≥100,000 colony-forming units of one type of bacterium were seen. 3) BC was deemed to be positive if non-contaminant bacteria were isolated from the blood.

85

4) CXR was positive if an abnormality such as pneumonia was visible.

RI PT

84

86

88

A positive UA that had no growth on culture (a negative UC) was defined as

SC

87

negative for a UTI.

We collected patient demographic data on gender, age, body mass index (BMI), and

90

Charlson comorbidity index (CCI). Additionally, perioperative records were collected,

91

including the type of procedure, hospital length of stay (LOS), and inpatient complications.

M AN U

89

With regards to POP, we investigated the rate of positive findings of these tests and

93

determined how many unnecessary diagnostic fever workups were performed. Then, we

94

assessed the clinical relevance of diagnostic workups in TJA patients with POP and the

95

relation of POP to febrile complications. We also investigated the relation of these

96

complications and in-hospital mortality to the pattern of POP and the type of procedure

97

performed.

99 100

EP

Finally, we investigated the estimated costs directly associated with diagnostic

AC C

98

TE D

92

workups of POP in TJA patients. Our institution charges $433.00 for each CXR (including radiologist fees), $26.40 for UA, $41.20 for UC, and $52.00 for BC.

101 102

Statistical Analysis

103

For comparison of demographics and POP characteristics according to the type of

104

TJA, Kruskal-Wallis tests or Fisher’s exact tests was used. The Fisher’s exact test was also 5

ACCEPTED MANUSCRIPT performed when comparing febrile complications and mortality according to the time and

106

extent of POP, and p-values were adjusted for three multiple comparisons using the Holm-

107

Bonferroni method. For comparison of the effects of POP characteristics and the type of TJA

108

on the positive rate of routine workups and complication rates, Kruskal-Wallis tests or

109

Fisher’s exact tests was used. Multivariate logistic regression was used to assess the risk of

110

positive diagnostic workup and febrile complication of POP, and a p-value <0.05 was

111

considered statistically significant. All statistical analyses were done with R software (version

112

3.11, R Foundation for Statistical Computing, Vienna, Austria).

SC

RI PT

105

114

M AN U

113

Source of funding:

115

No external funding was received for the preparation of this manuscript.

117

Results

118

Postoperative pyrexia

TE D

116

In our patient cohort, 11,875 patients (46.5%) developed POP during their hospital

120

stay, of whom 5,117 were males (43%) and 6,758 females (57%), with a mean age of 63.1

121

years (range, 16 to 98 years). The mean BMI was 30.3 kg/m2 (range, 14.1 to 69.8 kg/m2).

122

These patients underwent 5,596 (47.1%) primary total hip arthroplasties (THA), 4,665

123

(39.3%) primary total knee arthroplasties (TKA), 990 (8.3%) revision THAs, and 624 (5.3%)

124

revision TKAs. The mean CCI was 0.74 (range, 0 to 12) and the mean LOS was 3.9 days

125

(range, 1 to 192 days). In revision TJA patients, CCI was higher and LOS was longer than in

126

primary TJA (p<0.0001) (Table 1).

AC C

EP

119

127

The first episode of POP occurred on POD 1 for 41.2% of febrile patients and POD

128

2 for 46.2%. The average Tmax in primary and revision THA and TKA was the same

129

(101.3°F), and POP ≤102°F 24 hours after surgery occurred most commonly in all TJA 6

ACCEPTED MANUSCRIPT groups (p<0.0001). There were 7,893 patients (66.5%) who had multiple POP spikes. Of

131

these, 2,494 (31.6%) had two spikes, 1,600 (20.3%) had three spikes, and 1,096 patients

132

(13.9%) had four spikes of POP; and these multiple spikes were more common in revision

133

TKA patients (73.1%). POP >102°F was most frequently found in revision TKA patients

134

(18.6%). There was no POP >102°F in the single-spike group, but 19.5% of patients in the

135

multiple-spike group had POP >102°F (p<0.0001).

137

SC

136

RI PT

130

Routine diagnostic workup

Routine diagnostic workups for POP were performed in 90.5% (10,751) of all

139

patients with POP. These included 11,589 separate routine diagnostic tests: CXR (10,673),

140

UA (586), UC (201), and BC (129). Only 2.4% (274) of these tests were positive. UA showed

141

the highest positive rate (38.7%), followed by UC (9.5%), BC (7.0%), and CXR (0.18%).

M AN U

138

In patients with POP ≤102°F more than 24 hours after surgery, 95% (8,656/9,114)

143

underwent only CXR according to our protocol, and CXR was positive in very few cases

144

(0.12%). Meanwhile, in patients with POP >102°F, only 2.7% (42/1,536) underwent all

145

routine workups according to the protocol, and 13.1% (22/168) showed positive findings.

146

Among all patients with POP >102°F, 1,870 tests were performed with an overall positive

147

finding rate of 6.0%. CXR was only positive in 0.46% cases, compared with 43.4%, 11.4%,

148

and 11.3% positive findings in UA, UC, and BC, respectively (Table 2).

150

EP

AC C

149

TE D

142

Febrile complications and in-hospital mortality

151

Febrile complications occurred in 4.5% (531) of patients with POP. Patients with

152

POP >102°F had a higher complication rate (8.0%) and higher rate of infectious

153

complications (56.9%) than POP patients with temperatures ≤102°F (p<0.0001) (Table 3).

154

The most common complication was PE (1.5%), diagnosed by CT scan, and the incidence 7

ACCEPTED MANUSCRIPT was highest in patients with POP >102°F (2.1%). PJI was the second most common

156

complication in 1.2% of all patients with POP, and was highest in patients with POP >102°F

157

(2.2%). Meanwhile, pneumonia, UTI, and sepsis showed an overall low incidence compared

158

with other complications (0.16%, 0.16%, and 0.08%, respectively).

RI PT

155

The mortality rate of patients with POP was 0.13%. In-hospital mortality was the

160

highest in patients with POP >102°F (0.33%), although there were no statistically significant

161

differences among the different groups (p=0.18).

162

Factors associated with positive POP workups

M AN U

163

SC

159

The positive rate of fever workups significantly increased when the first POP

165

developed after POD 3 (p=0.0053). Tmax of POP >102°F (p<0.0001) and multiple fever

166

spikes (p=0.0008) resulted in higher rates of positive tests. Revision TJA showed a higher

167

positive rate of diagnostic tests than primary TJA (p<0.0001), even though the positive rate of

168

CXR showed no statistically significant difference between the two groups (Table 4).

TE D

164

Febrile complication rates were significantly higher in cases when the first POP

170

occurred after POD 3, POP was >102°F, multiple fever spikes, and revision TJA (p<0.0001)

171

(Table 5). The time of first POP (per 24-hour increase; odds ratio [OR] 1.29, 95% confidence

172

interval [CI] 1.18-1.42), the maximum temperature (per 1-degree increase; OR 1.43, 95% CI

173

1.28-1.60), multiple fever spikes (OR 1.66, 95% CI 1.30-2.13), and revision TJA (OR 5.29,

174

95% CI 3.99-7.01) were independent predictors of febrile complications according to

175

multivariate logistic regression analysis.

AC C

EP

169

176

177

Costs of routine workup for POP

8

ACCEPTED MANUSCRIPT A total of 11,589 tests that cost $4,652,239.40 were ordered for the diagnostic

179

workup of POP over a 12-year period. The most common cost was for CXR ($4,621,409.00;

180

98.9%). These costs did not include costs associated with an increased length of hospital stay

181

due to the evaluation or treatment of POP. The estimated cost for 11,319 negative workups

182

for POP was $4,636,976.80, with negative CXR costing $4,613,182.00 (99.2%).

183

Discussion

RI PT

178

POP following TJA is a common phenomenon in the early postoperative period [1].

185

There is evidence that POP in the early postoperative period occurs secondary to a normal

186

inflammatory response from the surgical insult, and the diagnostic workup for infection may

187

be unnecessary [10,17-22]. Despite this, concern regarding early detection of infection often

188

leads to an unnecessary diagnostic workup that can increase the cost of hospitalization for

189

patients with POP. Accordingly, this retrospective large-cohort study was conducted to

190

investigate the characteristics of POP in the early postoperative period after TJA to determine

191

the positive rates of commonly ordered tests and the cost-effectiveness of diagnostic workups,

192

and to evaluate the relationship between POP and febrile complications such as PJI. This has

193

led us to suggest a new evidence-based algorithm of managing POP after TJA.

EP

TE D

M AN U

SC

184

Czaplicki et al [6] reported that the incidence of pyrexia was highest on POD 1 in

195

THA patients and on POD 2 in TKA patients, which decreased after POD 2. Athanassious et

196

al [5] showed that greater than 50% of patients had a POP on POD 1 in both the THA and

197

TKA groups. We had similar findings; 87.4% of all POP following primary or revision TJA

198

developed on POD 1 or 2. However, POP was more likely to develop after POD 2 in revision

199

TJA, and these cases were associated with more febrile complications. Thus, more attention

200

should be paid to POP that develops after POD 2, particularly in the setting of a revision TJA.

AC C

194

9

ACCEPTED MANUSCRIPT To date, several studies have investigated the clinical significance of POP following

202

TJA [1-3,5-8,11,12,17-22]. Ghosh et al [1] examined 170 TKA patients and found 14

203

infections (8.2%); only 4 patients (28.6%) had a fever higher than 100.4°F. The authors

204

concluded that POP as a diagnostic indicator of infection has a low sensitivity and specificity,

205

and a positive predictive value. Similarly, Kennedy et al [21] found no association between

206

POP and infection in 92 TKA patients. In a study analyzing the development of POP after

207

THA, Than and Malovics [11] found that 39% of 177 patients had a fever exceeding 100.4°F,

208

but only 3 patients had an associated wound infection. In our study, POP occurred in 46% of

209

patients undergoing TJA, with a higher frequency than that reported in the literature [1-3,5-

210

7,9]. However, only 2.0% of patients with POP developed infectious complications such as

211

pneumonia, UTI, PJI, sepsis, or SSI. PJI occurred in only 1.2% of patients with POP.

212

Meanwhile, in patients with POP >102°F, both the infectious and febrile complication rates

213

were significantly greater compared to patients with POP ≤102°F. Therefore, there should be

214

a higher suspiction of febrile complications in patients with POP >102°F.

TE D

M AN U

SC

RI PT

201

In our study, the positive rate of routine diagnostic workups (2.4%) was much lower

216

than that reported in the literature. Several investigators have reported diagnostic rates for

217

individual tests that range between 11% and 26% [6,9,23,24]. Ward et al [9] reported a

218

positive rate of 14.8% of tests ordered in 161 patients who developed POP after TJA. The

219

positive rates of CXR, UA, UC, and BC were reported as 2%, 23.7%, 22%, and 6%,

220

respectively. Athanassious et al [5] reported positive CXR, UA, and UC rates of 4.3%, 19.1%,

221

and 0%, respectively, in 112 TJA patients. These studies concluded that CXR, UC, and BC

222

are not helpful in evaluating POP. Our study also showed that the positive rates of CXR

223

(0.18%), UC (9.5%), and BC (7.0%) were low, while the positive rate of UA (38.7%) was

224

high. The positive rate of CXR was very low compared with those reported in other studies.

AC C

EP

215

10

ACCEPTED MANUSCRIPT 225

This may be due to the fact that a much larger number of patients with POP (89.9%) in our

226

cohort had a CXR done as part of a routine workup. The extremely low rate of positive CXR

227

suggests that CXR should not be part of a routine diagnostic workup for POP. Ward et al [9] noted that a fever that develops after POD 3 and multiple days of

229

fever are independent predictors of a positive evaluation, and that a Tmax >102°F increased

230

the likelihood of a positive evaluation. Similarly, the positive rate of routine workups in our

231

study significantly increased when POP developed after POD 3, occurred after revision TJA,

232

if there were multiple fever spikes, or when POP was >102°F. However, routinely performing

233

all fever diagnostic tests in this group is also not warranted. A thorough assessment of clinical

234

symptoms should be used to guide clinicians in deciding the type and order of tests to be

235

performed. An evidence-based, new algorithm of treating POP (Figure2) is suggested to

236

eliminate unnecessary diagnostic workups, minimize cost, and most importantly, prevent

237

patients from undergoing potentially invasive tests.

TE D

M AN U

SC

RI PT

228

Several authors have assessed the cost-effectiveness of diagnostic workups

239

[9,23,25]. Ward et al [9] reported that the cost of all POP evaluations for patients undergoing

240

TJA for a two-year period was $73,878; on the other hand, the overall cost per positive POP

241

evaluation was $3,358. In our study, the total estimated cost of routine workups for POP after

242

TJAs during the 12-year period was $4,652,239.40, and the total estimated cost for 11,319

243

negative tests in patients with POP was $4,636,976.80, the largest of which was the cost

244

associated with CXR ($4,613,182). If these unnecessary workups were not performed,

245

$386,414.73 would have been saved per year. The economic burden of these tests is even

246

more significant because these tests may prolong hospital stay while waiting for laboratory

247

confirmation of culture results, and the cost of sample collection, laboratory processing, and

248

imaging increases. Our study, however, could not determine the cost of delayed hospital

AC C

EP

238

11

ACCEPTED MANUSCRIPT 249

discharge for this cohort. The current study has several limitations. It is a retrospective chart review and there

251

is variability in ordering and charting practices. Second, it was not noted which patients had a

252

blood transfusion, when the transfusion was performed, and if a transfusion reaction was

253

documented, as blood transfusion can lead to POP. Third, body temperature was measured

254

orally, which is not a true measurement of core temperature. However, all patients were

255

measured in the same way. Fourth, all patients received perioperative antibiotics up to POD 1

256

and antibiotics may have blunted the febrile response. Fifth, in the last four years of the study,

257

a multimodal pain control protocol was instituted that includes the use of perioperative

258

acetaminophen, which has been shown to decrease the amount of POP with a consequent

259

decrease in POP workup [25]. Finally, no predictions of POP characteristics and positive

260

diagnostic tests using multivariate logistic regression could be performed because the positive

261

rates of routine workups in each subgroup were too low.

TE D

M AN U

SC

RI PT

250

To our knowledge, the present study provides the most comprehensive and largest

263

analysis of POP workup after TJA. Although pyrexia is a common occurrence after TJA,

264

routine workups without taking symptoms into account are neither clinically useful nor cost-

265

effective for evaluating early-onset POP. Therefore, good clinical assessment of definite

266

causes of POP, as well as a new, evidence-based algorithm for management of POP are

267

needed to minimize variations in practice and limit diagnostic workups that are unnecessarily

268

costly.

AC C

EP

262

269

12

ACCEPTED MANUSCRIPT References

271 272

1. Ghosh S, Charity RM, Haidar SG, Singh BK. Pyrexia following total knee replacement. The Knee. 2006 Aug;13(4):324–7.

273 274

2. Summersell PC, Turnbull A, Long G, Diwan A, Macdessi S, Cooke PJ, Li GS. Temperature trends in total hip arthroplasty: a retrospective study. J Arthroplasty. 2003 Jun;18(4):426–9.

275 276

3. Tai TW, Chang CW, Lin CJ, Lai KA, Yang CY. Elevated temperature trends after total knee arthroplasty. Orthopedics. 2009 Dec;32(12):886.

277 278

4. Engoren M. Lack of association between atelectasis and fever. Chest. 1995 Jan;107(1):81– 4.

279 280 281

5. Athanassious C, Samad A, Avery A, Cohen J, Chalnick D. Evaluation of fever in the immediate postoperative period in patients who underwent total joint arthroplasty. JArthroplasty. 2011 Dec;26(8):1404–8.

282 283 284

6. Czaplicki AP, Borger JE, Politi JR, Chambers BT, Taylor BC. Evaluation of postoperative fever and leukocytosis in patients after total hip and knee arthroplasty. J Arthroplasty. 2011 Dec;26(8):1387–9.

285 286

7. Guinn S, Castro FP, Garcia R, Barrack RL. Fever following total knee arthroplasty. Am J Knee Surg. 1999 Summer;12(3):161–4.

287 288

8. Vijaysegaran P, Coulter SA, Coulter C, Crawford RW. Blood cultures for assessment of postoperative fever in arthroplasty patients. J Arthroplasty. 2012 Mar;27(3):375–7. 5.

289 290 291

9. Ward DT, Hansen EN, Takemoto SK, Bozic KJ. Cost and effectiveness of postoperative fever diagnostic evaluation in total joint arthroplasty patients. J Arthroplasty. 2010 Sep;25(6Suppl):43–8.

292 293

10. Andres BM, Taub DD, Gurkan I, Wenz JF. Postoperative fever after total knee arthroplasty: the role of cytokines. Clin Orthop Relat Res. 2003 Oct;415:221-31.

294 295

11. Than P, Málovics I. [Significance of postoperative fever after hip prosthesis implantation]. Z Orthop Ihre Grenzgeb. 2000 Sep-Oct;138(5):430–5.

296 297

12. Shaw JA, Chung R. Febrile response after knee and hip arthroplasty. Clin Orthop Relat Res. 1999 Oct;367:181-9.

298 299 300 301

13. Parvizi J, Zmistowski B, Berbari EF, Bauer TW, Springer BD, Della Valle CJ, Garvin KL, Mont MA, Wongworawat MD, Zalavras CG. New definition for periprosthetic joint infection: from the Workgroup of the Musculoskeletal Infection Society. Clin Orthop Relat Res. 2011 Nov;469(11):2992-4.

302 303 304

14. Schwartz DS, Barone JE. Correlation of urinalysis and dipstick results with catheterassociated urinary tract infections in surgical ICU patients. Intensive Care Med. 2006 Nov;32(11):1797–801.

AC C

EP

TE D

M AN U

SC

RI PT

270

13

ACCEPTED MANUSCRIPT 15. Dos Santos JC, Weber LP, Perez LRR. Evaluation of urinalysis parameters to predict urinary-tract infection. Braz J Infect Dis. 2007 Oct;11(5):479–81.

307 308

16. Tambyah PA. Catheter-associated urinary tract infections: diagnosis and prophylaxis. Int J Antimicrob Agents. 2004 Sep;24 Suppl 1:S44–8.

309 310

17. Bindelglass DF, Pellegrino J. The role of blood cultures in the acute evaluation of postoperative fever in arthroplasty patients. J Arthroplasty. 2007 Aug;22(5):701–2.

311 312 313

18. Cartmell T, Poole S, Turnbull AV, Rothwell NJ, Luheshi GN. Circulating interleukin-6 mediates the febrile response to localised inflammation in rats. J Physiol. 2000 Aug;526 Pt 3:653–61.

314 315 316

19. Hahn HH, Char DC, Postel WB, Wood WB. Studies on the pathogenesis of fever. XV. The production of endogenous pyrogen by peritoneal macrophages. J Exp Med. 1967 Aug;126(2):385–94.

317 318 319

20. Handel M, Winkler J, Hörnlein RF, Northoff H, Heeq P, Teschner M, Sell S. Increased interleukin-6 in collected drainage blood after total knee arthroplasty: an association with febrile reactions during retransfusion. Acta Orthop Scand. 2001 Jun;72(3):270–2.

320 321 322

21. Kennedy JG, Rodgers WB, Zurakowski D, Sullivan R, Griffin D, Beardsley W, Sheehan L. Pyrexia after total knee replacement. A cause for concern? Am J Orthop (Belle Mead NJ). 1997 Aug;26(8):549-52.

323 324

22. Fanning J, Brewer J. Delay of hospital discharge secondary to postoperative fever--is it necessary? J Am Osteopath Assoc. 2002 Dec;102(12):660–1.

325 326 327

23. Da Luz Moreira A, Vogel JD, Kalady MF, Hammel J, Fazio VW. Fever evaluations after colorectal surgery: identification of risk factors that increase yield and decrease cost. Dis Colon Rectum. 2008 May;51(5):508–13.

328 329

24. Schey D, Salom EM, Papadia A, Penalver M. Extensive fever workup produces low yield in determining infectious etiology. Am J Obstet Gynecol. 2005 May;192(5):1729–34.

330 331 332 333

25. Kocher MS, Erens G, Thornhill TS, Ready JE. Cost and effectiveness of routine pathological examination of operative specimens obtained during primary total hip and knee replacement in patients with osteoarthritis. J Bone Joint Surg Am. 2000 Nov;82-A(11):1531– 5.

334 335 336

26. Karam JA, Zmistowski B, Restrepo C, Hozack WJ, Parvizi J. Fewer postoperative fevers: an unexpected benefit of multimodal pain management? Clin Orthop Relat Res. 2014 May;472(5):1489–95.

AC C

EP

TE D

M AN U

SC

RI PT

305 306

337

338

14

ACCEPTED MANUSCRIPT Table 1 Patient and Procedural Demographics. Total

Primary THA

Primary TKA

Revision THA

Revision TKA

(%)

(%)

(%)

(%)

(%)

AC C

EP

TE D

M AN U

SC

RI PT

Total (number) 11875 5596 4665 990 624 Age (years) ≤ 50 14.2 20.6 6.8 16.4 8.5 51 – 70 59.9 55.4 66.8 49.7 64.4 > 70 25.9 24.0 26.4 36.9 27.1 2 BMI (kg/m ) ≤ 30 52.8 61.1 41.9 60.0 40.5 31 – 40 36.1 29.9 43.4 30.8 46.5 > 40 11.1 9.0 14.7 9.2 13.0 Length of stay (days) ≤3 62.5 71.8 58.9 41.0 40.1 4–6 29.9 22.5 36.1 36.1 39.6 ≥7 7.6 5.7 5.0 22.9 20.3 CCI (average ± SD) 0.74 ± 1.04 0.62 ± 0.96 0.80 ±1.00 0.88 ±1.30 1.06 ±1.24 Pyrexia Tmax (average ± SD) 101.3 ± 0.71 101.3 ± 0.67 101.3 ± 0.71 101.3 ± 0.91 101.3 ± 0.75 ≥100.4° F to ≤102° F 87.0 87.3 87.5 86.8 81.4 ≤24 hours 10.3 12.0 9.8 5.5 5.9 >24 hours 76.7 75.3 77.7 81.3 75.5 > 102° F at any time 13.0 12.7 12.5 13.2 18.6 Frequency of pyrexia Single event 33.5 33.4 34.8 32.4 26.9 Multiple events 66.5 66.6 65.2 67.6 73.1 THA, total hip arthroplasty; TKA, total knee arthroplasty; BMI, body mass index; CCI, Charlson comorbidity index; SD, standard deviation; Tmax, maximum temperature; F=Farenheight.

p-value

< 0.0001

< 0.0001

< 0.0001 < 0.0001

< 0.0001

0.001

ACCEPTED MANUSCRIPT Table 2 Positive Findings of Routine Diagnostic Workups in Total Joint Arthroplasty Patients with Postoperative Pyrexia. N = 11,875 ≥100.4° F to ≤102° F ≤ 24 hours postoperative

3.2% (1/31)

UA

15.1% (11/73)

UC

7.1% (2/28)

BC

0% (0/6) 10.1% (14/138)

CXR

0.12% (11/9,107)

UA

41.2% (124/301)

UC

8.7% (9/103)

BC

4.3% (3/70)

1.5% (147/9,581)

>102° F at any time

EP

CXR

TE D

Total positive tests

UA

AC C

UC BC

n = 9,114

M AN U

≥100.4° F to ≤102° F > 24 hours postoperative

SC

Total positive tests

Total positive tests

n = 1,536 0.46% (7/1,535) 43.4% (92/212) 11.4% (8/70) 11.3% (6/53) 6.0% (113/1,870)

F, Fahrenheit; CXR, chest X-ray; UA, urinalysis; UC, urine culture; BC, blood culture.

RI PT

CXR

n = 1,125

ACCEPTED MANUSCRIPT Table 3 Complications and Mortality of Patients with Postoperative Pyrexia according to Time and Extent. Complications

> 102° F at any time (n=1536) (Group 3)

48.1% (n=13)

41.2% (n=157)

56.9% (n=70)

Pneumonia

1

11

7

UTI

2

9

8

PJI

8

104

34

Sepsis

0

3

6

SSI

2

30

15

51.9% (n=14)

58.8% (n=224)

43.1% (n=53)

PE

8

137

32

DVT

6

87

Total

2.2% (n=27)

In-hospital mortality

0.16% (n=2)

4.2% (n=381)

RI PT

SC

Non-infection

0.10% (n=9)

p-value

Group 1 vs Group 2: 0.11

21

8.0% (n=123)

M AN U

Infection

≥100.4° F to ≤102° F ≤24 hours (n=1225) >24 hours (n=9114) (Group 1) (Group 2)

0.33% (n=5)

Group 1 vs Group 3: < 0.0001 Group 2 vs Group 3: < 0.00001

Group 1 vs Group 2: 0.011 Group 1 vs Group 3: 0.00045 Group 2 vs Group 3: 0.03 Group 1 vs Group 2: 0.001 Group 1 vs Group 3: <0.0001 Group 2 vs Group 3: <0.0001 Group 1 vs Group 2: 1.0 Group 1 vs Group 3: 1.0 Group 2 vs Group 3: 0.18

P-values adjusted for 3 multiple comparisons (Holm-Bonferroni method).

AC C

EP

TE D

F, Fahrenheit; UTI, urinary tract infection; PJI, periprosthetic joint infection; SSI, surgical site infection; PE, pulmonary embolism; DVT, deep venous thrombosis.

ACCEPTED MANUSCRIPT Table 4 Positive Routine Diagnostic Workups according to Pyrexia Characteristics and the Type of TJA.

Frequency of POP Single (n=3982) Multiple (n=7893)

BC (%)

0.16

30.9

5.6

6.2

0

62.0

16.7

11.8

0.09

29.7

3.8

4.0

0.55

43.6

10.4

11.1

0.07

32.3

8.6

0

0.22

41.2

9.6

Total tests (%)

p-value

2.3 (253/11117) 4.4 (21/472)

0.0053

1.6 (152/9424) 5.6 (122/2165)

< 0.0001

1.8 (58/3264) 2.6 (244/8325)

RI PT

UC (%)

SC

Degrees of Tmax 100.4 – 101.9° F (n=10047) ≥ 102° F (n=1828)

UA (%)

M AN U

Time of First POP POD ≤ 3 (n=11453) POD > 3 (n=422)

CXR (%)

9.0

0.0008

Type of TJA 1.9 Primary 0.17 36.5 8.5 5.0 (187/9765) (n=10261) 4.8 < 0.0001 Revision 0.20 44.9 11.1 8.7 (n=1614) (87/1824) CXR, chest X-Ray; UA, urinalysis; UC, urine culture; BC, blood culture;

AC C

EP

TJA, total joint arthroplasty.

TE D

POP, postoperative pyrexia; Tmax, maximum temperature; F, Farenheight;

ACCEPTED MANUSCRIPT Table 5 Febrile Complications according to the Pyrexia Characteristics and the Type of TJA. Febrile Complication (%)

p-value

OR (95% CI)

p-value

Time of First POP 3.9

POD 2-3

4.2

POD > 3

13.7

<0.0001

Per 24-hour increase

RI PT

POD 0-1

1.29 (1.18-1.42)

Degrees of Tmax 3.5

101.3° – 102°F

4.6

> 102°F

7.8

<0.0001

SC

100.4° – 101.2°F

Per 1-degree increase

1.43 (1.28-1.60)

Single

2.8

Multiple

5.3

Type of TJA Primary

3.1

Revision

13.0

M AN U

Frequency of POP

<0.0001

<0.0001

<0.0001

1.66 (1.30-2.13)

<0.0001

5.29 (3.99-7.01)

<0.0001

<0.0001

AC C

EP

TJA, total joint arthroplasty.

TE D

POP, postoperative pyrexia; POD, postoperative day; Tmax, maximum temperature; F, Farenheight;

ACCEPTED MANUSCRIPT 1

Figure Legend:

2

Figure 1.

Current algorithm for treatment of postoperative pyrexia at our institution.

Figure 2.

New proposed algorithm for treatment of postoperative pyrexia.

4

RI PT

3

AC C

EP

TE D

M AN U

SC

5

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT