Intraoperative evaluation of polymorphonuclear leukocyte during second-stage revision surgery promote overdiagnosis of persistent periprosthetic joint infection

Abstract

Objective

The aim of this study was to evaluate whether intraoperative histopathological examination could predict the risk of relapse of infection in periprosthetic joint infections (PJI).

Methods

The study included 25 patients (14 women and 11 men, with a mean age of 67.0 years (range, 37–83 years)), who had two-staged revision surgery for a PJI. Following prosthetic removal in the first stage, all patient underwent an intraoperative histopathological examination during the second stage. The patients were divided into PMNs-positive group (≥five PMNs per high-powered field) or -negative group (<five PMNs). A relapse was defined as the occurrence of PJI. Median follow-up was 51 months (range, 32–80 months) following second-stage revision surgery.

Results

Intraoperative histopathological revealed that 8.0% of cases were PMNs-positive. Postoperative histopathological examination revealed that 28.0% of cases were PMNs-positive. 28.0% of cases showed discrepancy between the PMNs-positivity. Intraclass correlation coefficient indicates poor reproducibility. Infection relapse after revision surgery occurred in two cases (8.0%); both relapse cases were from the PMNs-negative group. There was no statistical relationship between the presence of PMNs in periprosthetic tissue by intraoperative or postoperative histopathological examination and relapse of infection.

Conclusions

Our findings showed that intraoperative histopathological examination could not predict the relapse of infection. Intraoperative histopathological examination promotes overdiagnosis of the requirement for re-implantation of antibiotic-impregnated cement and prolonged treatment periods.

Level of evidence

Level III, diagnostic study

Introduction

Periprosthetic joint infection (PJI) is one of the most common and intractable complications following total joint arthroplasty.¹ PJI is treated with either a one-stage exchange operation or a two-stage exchange using an antibiotic-impregnated cement spacer.2, 3 When the infection is not abrogated, re-exchange of antibiotic-impregnated cement spacer is conducted. Infection persistence is diagnosed by a combination of clinical presentation, blood tests, culture, and histological examinations.4, 5 Musculoskeletal Infection Society reported that accumulation of polymorphonuclear leukocytes (PMNs) in tissues is a minor criterion for diagnosis of persistent infection.⁵ In addition, intraoperative histopathological examination is reportedly useful for diagnosing persistent infection during second-stage revision surgery.6, 7, 8 However, we experienced several cases with discrepancy between the PMNs-positivity of intraoperative frozen sections compared with postoperative hematoxylin and eosin (HE)-stained sections in second-stage revision surgery. These PMNs positive cases did not relapse PJI after second-stage revision surgery. We hypothesized that although intraoperative PMNs-positivity can detect infection persistence,7, 9, 10 PMNs-positivity cannot predict the relapse of PJI after revision arthroplasty following first stage antibiotic-impregnated cement spacer replacement. Therefore, we retrospectively compared the PMNs-positivity in periprosthetic tissues by examining intraoperative or postoperative HE sections, and compared the results of clinical follow-up with these histopathological results.

Methods

Patients

This study is a single center retrospective study. This study was based on the data of 25 patients who had been hospitalized in the Department of Orthopaedic Surgery of Kagoshima University hospital from October 2006 to October 2014. All of two stage revision surgery patients who were performed intraoperative histopathological examinations were included in this study. Prosthesis removal for treatment of PJI was done in 8 bipolar hip arthroplasty (BHA) patients, 10 total hip arthroplasty (THA) patients, and 7 total knee arthroplasty (TKA) patients. There were 14 women and 11 men, with a mean age of 67.0 years (range, 37–83 years). Before removal of the prosthesis or after revision arthroplasty, a diagnosis of PJI was made based on clinical presentation, laboratory data, and diagnostic imaging, as recommended by the Infectious Diseases Society of America.8, 11

First-stage surgery involved prosthesis removal, debridement, irrigation, and placement of antibiotic-impregnated spacer. Intravenous antibiotics were administered for 4–6 weeks following first-stage surgery. The abrogation of PJI was assessed after a 2-week drug-free period. Second-stage revision surgery was indicated in patients who did not demonstrate local heat, tenderness, elevation of C-reactive protein (CRP), and elevation of erythrocyte sedimentation rate. Following a 2–3 months placement of antibiotic-impregnated spacer period, second-stage revision surgery was performed. Median follow-up was 51 months (range, 32–80 months) following second second-stage revision surgery.

Intra- and post-operative histopathological examinations

The cement spacer was impregnated with 2.5–10% vancomycin combined with 0–5.0% amikacin sulfate. A cement spacer mold (Biomet Orthopedics, USA) or handmade cement spacer was used. Following prosthesis removal, intravenous antibiotics were administered for 4–6 weeks. When methicillin-resistant Staphylococcus or methicillin-resistant Staphylococcus aureus (MRSA) was detected, vancomycin or teicoplanin were used; cefazolin was used for all other cases after drug susceptibility testing. Antibiotics were intravenously administered twice daily. All patients were allowed to use a wheelchair and to move their affected joints between operations. Hip and knee joints were allowed mobilization. Between three and seven periprosthetic specimens (median, three) were collected from each patient during each second-stage revision surgery; intraoperative, postoperative histopathological evaluation and bacterial culture was performed on these specimens using blood agar medium and BTB agar. Culture duration was 7 days. Detection of bacterial strain was performed by mass spectrometry using VITEK MS (BioMérieux, Lyon, France). Previous studies reported that five PMNs per high-powered field (HPF) was a suitable diagnostic threshold for diagnosis of PJI¹²; therefore, we regarded cases with five PMNs per HPF in more than five fields as PJI positive. We divided the patients into the PMNs-positive group (≥five PMNs per HPF) and the PMNs-negative group (<five PMNs per HPF). Manifestation of PJI following second-stage revision surgery was identified as a relapse of infection. All histopathological examinations were performed by skilled pathologists who were blinded to the grouping data.

Statistical analysis

The intraclass correlation coefficient assessment was performed to examine the reproducibility between intra- and postoperative histopathological examinations. Fisher's exact test was performed to evaluate the differences between groups. All statistical analyses were performed using BellCurve for Excel 2015 (Social Survey Research Information Co. (SSRI), Ltd, Osaka, Japan) which is an add-in software for Excel (Microsoft, Redmond, USA).

Results

The intraclass correlation coefficient indicates very poor reproducibility of PMNs-positivity between intraoperative and postoperative histopathological examinations.

The individualized data are outlined in Table 1. Intraoperative histopathological examination of frozen sections of tissue obtained during second-stage revision surgery revealed that 8.0% of cases (2/25) were PMNs-positive. In contrast, postoperative histopathological examinations of HE sections revealed that 28.0% of cases (7/25) were PMNs-positive (Table 1, Table 2). Seven cases (28.0%) showed discrepancy between the results of intraoperative and postoperative HE histopathological examinations (red characters in Table 1). The intraclass correlation coefficient indicate very poor reproducibility of intraoperative and postoperative histopathological examinations (0.116: −0.233–0.462) (p = 0.267) (Table 1). In addition, Fisher's exact test showed that there was no statistical relationship between PMNs-positivity in intra- and postoperative histopathological examinations (Table 2).

Table 1.

Clinicopathological data of patients with periprosthetic joint infection.

SEX	age	op	bacteria for 1st op	PMNs in intraop	PMNs in permanent	culture (second stage)	comorbidity	DM	BMI	WBC	Hb	CRP	follow-up (days)	Relapse
F	69	reTHA	MRCNS	<5	<5	–	RA		22.4	4350	11.2	0.10	2917
F	74	BHA	Unspecified	<5	<5	–	–		22.9	4090	12.5	0.27	2400
M	70	BHA	MRCNS	<5	<5	–	–		22.4	4390	13.4	0.12	3005
F	72	THA	MRSA	<5	<5	–	ITP		23.2	3700	11.6	0.65	3304
F	58	BHA	Unspecified	<5	<5	CNS			19.8	4000	12.7	0.41	3636
M	37	BHA	MRCNS	<5	<5	–			18.2	3910	13.1	0.17	952
F	66	BHA	MRCNS	<5	<5	–	MC	+	17.4	2920	11	0.02	937
M	68	THA	MRSA	<5	<5	–			19.3	4190	11.8	0.12	958
F	72	THA	Unspecified	<5	<5	CNS			20.6	3460	10.3	0.12	914
M	77	TKA	Unspecified	<5	<5	–			24.2	4640	11.8	0.23	1086
F	72	TKA	MSSA	<5	<5	–			28.4	7020	12.6	0.14	2079
F	66	TKA	MSSA	<5	<5	–			23.1	4160	11.3	0.24	2889
F	74	THA	Staphylococcus aureus	<5	<5	–		+	26.9	4350	11.0	0.09	1093
M	77	THA	Unspecified	<5	<5	–	AF		27.9	3860	11.7	0.14	926
F	83	TKA	Unspecified	≥5	<5	–			20.9	4230	11.2	0.67	1803
M	78	TKA	Streptococcus agalactiae	<5	<5	–	–		19.3	3900	10.9	0.23	1032
M	65	reTHA	unspecified	<5	≥5	–	CI		30.2	3670	13.6	0.11	1824
F	57	THA	unspecified	<5	≥5	–	–	+	24.4	3860	12.5	0.15	1987
M	54	BHA	unspecified	<5	≥5	–	CI	+	21.0	5930	13.3	0.02	1025
M	52	BHA	unspecified	<5	≥5	–	drug addict		24.9	4800	10.7	0.01	884
F	70	BHA	unspecified	<5	≥5	–	-	+	21.8	5280	11.6	0.11	1063
F	66	THA	unspecified	<5	≥5	–	RA		14.4	4020	10.3	3.62	777
F	78	TKA	MSSA	≥5	≥5	–			24.4	3010	12.3	0.09	1406
M	70	THA	unspecified	<5	<5	–	CI		22.0	4790	12.9	0.05	1219	+
M	51	TKA	MRCNS	<5	<5	MRSA			25.8	6820	12.5	0.06	2219	+

reTHA: revision THA, BHA: bipolar hip arthroplasty, THA: total hip arthroplasty, TKA: total knee arthroplasty, MRCNS: methicillin-resistant coagulase negative staphylococcus, MRSA: methicillin-resistant Staphylococcus aureus, MSSA: methicillin-sensitive Staphylococcus aureus, CNS: coagulase-negative Staphylococcus, PMNs: polymorphonuclear leukocytes, ITP: idiopathic thrombocytopenic purpura, MC: multiple cancer, RA: rheumatoid arthritis, AF: atrial fibrillation, CI: cerebral infarction, DM: diabetes mellitus, BMI: body mass index, Hb: hemoglobin, WBC: white blood cell count/ml, CRP: C-reactive protein (mg/dL).

Table 2.

Number of cases in each group that showed PMNs-positive results.

	Postoperative histopathological examination: PMNs (+)	Postoperative histopathological examination: PMNs (−)
Intraoperative histopathological examination: PMNs (+)	1	1
Intraoperative histopathological examination: PMNs (−)	6	17

PMNs: polymorphonuclear leukocytes. (p = 0.490).

There was no statistical relationship between accumulation of PMNs in intraoperative or postoperative histopathological examinations and relapse of infection.

The proportion of infection relapse was 8.0% (2/25). In one case of relapse, PJI was abrogated by third-stage irrigation and debridement; the other case of relapse had the revision THA removed. All cases of infection relapse were from the PMNs-negative group as defined by intra- and postoperative histopathological findings. Fisher's exact test revealed no statistical relationship between PMNs-positivity in intra- or postoperative histopathological examinations and relapse of infection (Table 3).

Table 3.

Number of cases in each group that showed relapse of infection.

		Relapse of infection −	Relapse of infection +	p-value
Intraoperative histopathological examination	PMNs-negative	21	2	p = 1.000
	PMNs-positive	2	0
Postoperative histopathological examination	PMNs-negative	16	2	p = 1.000
	PMNs-positive	7	0

PMNs: polymorphonuclear leukocytes.

There was no statistical relationship between accumulation of PMNs in intraoperative or postoperative histopathological examinations and detection of bacteria.

Pathogenic bacteria were detected in samples obtained during second-stage revision surgery in three cases, including two cases of coagulase-negative staphylococci and one of methicillin-resistant Staphylococcus aureus. The causative agent in one case of infection relapse was detected as methicillin-resistant coagulase-negative Staphylococcus (MRCNS) in a sample obtained during second-stage revision surgery, while there was no specific bacteria detected in the other infection relapse case. The three cases in which bacteria were detected were from the PMNs-negative group by intra- and postoperative histopathological examination. Fisher's exact test revealed no statistical relationship between PMNs-positivity and positive culture results of revision samples (Table 4).

Table 4.

Bacterial culture results in each group.

		Culture-negative	Culture-positive	p-value
Intraoperative histopathological examination	PMNs-negative	20	3	p = 1.000
	PMNs-positive	2	0
Postoperative histopathological examination	PMNs-negative	15	3	p = 0.534
	PMNs-positive	7	0

PMNs: polymorphonuclear leukocytes.

Discussion

We showed that intraclass correlation coefficient indicates very poor reproducibility of PMNs-positivity between intraoperative and postoperative histopathological examinations; although these discrepancies might be caused by technical difficulties associated with intraoperative examination of frozen sections or subtle differences in microscopic examination area, the results of postoperative histopathological examinations evaluation should provide more accurate results. We also found no statistical relationship between PMNs-positivity in intraoperative or postoperative histopathological examinations and relapse of infection. These findings indicate that the significance of intraoperative histopathological examinations during second-stage revision surgery following removal of prosthesis and implantation of antibiotic-impregnated cement spacer is questionable.

Two-stage revision surgery has been regarded as the gold standard for treatment of PJI; however, one-stage revision surgery is also a valid treatment choice. Several systematic reviews showed similar rates of infection relapse between one- or two-stage revisions for PJI treatment.13, 14, 15, 16 Single-stage revision surgery for PJI requires aggressive debridement of all necrotic tissue and pulsatile lavage with saline¹⁷; this decreases the number of bacteria and purulent or ischemic tissues. First-stage operation of two-stage revision decreased the number of bacteria, purulent tissue, and ischemic tissues, as shown by the reduction of CRP and erythrocyte sedimentation rate. In addition, the irrigation and debridement during first stage and second-stage surgery were equivalent to those of during one-stage prosthesis exchange. Although, evaluation of postoperative histopathological examinations showed that 28.0% of 25 revision cases had an accumulation of ≥5 PMNs in periprosthetic tissues, we showed no statistical relationship between PMNs-positivity in intraoperative or postoperative histopathological examinations and relapse of infection. These findings indicate that the severity of infection during second-stage revision following antibiotic-impregnated cement spacer placement might be similar to good prognosis cases by one-stage revision. Other possibility is that intra- and post-operative histopathological examinations cannot define the persistence of infection. Both hypotheses indicate that intraoperative histopathological examinations are not needed during second-stage revision surgery following removal of prosthesis and implantation of antibiotic-impregnated cement spacer. When intraoperative histopathological examinations are positive for PMNs, re-implantation of antibiotic-impregnated cement spacer is recommended to abrogate PJI. Our findings suggest that intraoperative histopathological examinations examination promote overdiagnosis of the requirement for re-implantation of antibiotic-impregnated cement.

This research has several limitations. The small sample size decreased the power of the study. Because this research is retrospective research, preoperative treatment, the number of samples during second-stage revision surgery, duration of cement spacer, duration of anti-biotics treatment following cement implantation, and dose of antibiotics in cement were divergent. Relapse of infection and de novo infection following second stage revision surgery was not distinguishable. One relapse case detected different bacterial strain from MRCNS to MRSA. This change from MRCNS to MRSA might be caused by microbial substitution or de novo infection.

Conclusions

Although our findings showed that intra- and postoperative histopathological examination could not predict the relapse of infection, removal of prosthesis and treatments of infection achieved a success rate of 92% (23/25) of revision without relapse. Examination of intraoperative histopathological examinations promote overdiagnosis for the necessity of prolonged treatment periods by re-implantation of antibiotic-impregnated cement.

Declarations

Ethics approval and consent to participate

This research protocol was approved by the Ethics Committee on Clinical Research at Kagoshima University Hospital (Pathological examination of periprosthetic infection for arthroplasty: No. 493). This study was conducted in accordance with the Helsinki Declaration of 1975, as revised in 2000. All patients gave their informed written consent for participation in this clinical study.

Consent for publication

All patients gave their informed written consent for participation in this clinical study.

Competing interests

The authors declare no financial or non-financial competing interests.

Funding

No funding supported.

List of abbreviations

BHA

bipolar hip hemiarthroplasty

CRP

C-reactive protein

HE

hematoxylin and eosin

HPF

high-powered field

PJI

periprosthetic joint infection

PMNs

polymorphonuclear leukocytes

THA

total hip arthroplasty

TKA

total knee arthroplasty

MRCNS

methicillin-resistant coagulase-negative Staphylococcus