Aggressive Early Débridement for Treatment of Acutely Infected Cemented Total Hip Arthroplasty

Abstract

Background

Up to 2% of THAs are complicated by infection, leading to dissatisfied patients with poor function and major social and economic consequences. The challenges are control of infection, restoration of full function, and prevention of recurrence. Irrigation and débridement with or without exchange of modular components remains an attractive alternative to two-stage reimplantation in acutely infected THAs but with variable results from previous studies.

Questions/Purposes

We therefore determined the rate at which aggressive early débridement with exchange of modular components controlled acutely infected cemented THAs.

Methods

We retrospectively studied all 26 patients presenting with acutely infected cemented THAs (16 primaries, 10 revisions) occurring within 6 weeks of the index operation or of hematogenous spread from a confirmed source of infection elsewhere between 1999 and 2006. Microbiology confirmed bacterial colonization in all patients, with 18 early postoperative infections and eight acute hematogenous infections. Patients underwent aggressive débridement with open complete synovectomy, exchange of all mobile parts, débridement of all aspects of the joint, irrigation with antiseptic solutions, and pulsatile lavage. Minimum followup was 5 years (mean, 6.6 years; range, 5–11 years).

Results

Eight patients had multiple washouts with control of infection in four at latest followup. Five patients underwent two-stage revisions and one patient was placed on long-term antibiotic suppression. Twenty patients returned to their expected functional level with no radiographic evidence of prosthetic failure. At minimum 5-year followup, we had a 77% infection control rate.

Conclusions

Our data confirm current literature and suggest there may be a role for aggressive early débridement in controlling acute postoperative and hematogenous infections after cemented THA.

Level of Evidence

Level IV, therapeutic study. See Instructions for Authors for a complete description of levels of evidence.

Introduction

Periprosthetic infection after THA was one of the most common complications in the early years of THA, occurring in 9.5% of patients in one series [34], and in the majority of cases presumably due to contamination during surgery [9]. More recently, precautions to prevent infection, such as the use of prophylactic antibiotics, ultraclean air in operating rooms, and improved surgical technique, have contributed to the reduction of periprosthetic infections [19]. However, despite the rates of infection falling to less than 2% of all primary THAs and less than 5% of revision THAs [36], the number of THAs has increased due to the aging population.

Tsukayama et al. [39] classified infections after THA into four categories: (1) positive intraoperative cultures from revision THA without other features of obvious infection; (2) early postoperative infections occurring less than 1 month postoperatively; (3) late chronic infections occurring more than 1 month postoperatively with an insidious onset; and (4) acute hematogenous infections where the infection occurs after a salient period of time from the primary procedure whereby the prosthesis becomes infected from hematogenous seeding.

The diagnosis of infection after THA can be challenging because no investigation is 100% sensitive and specific [14]. Therefore, a combination of clinical assessment, serologic tests, and imaging, as well as hip aspirates and biopsies, have been used to confirm the diagnosis [36].

Management of such infections is also difficult, often requiring a prolonged course of treatment resulting in increased cost to the healthcare system [15] and leading to dissatisfied patients with poor function. Treatment options available include débridement with retention of components [3, 6, 17, 19], single-stage revision [20, 36], two-stage revision and multistage revision [24], long-term suppressive antibiotics, and salvage procedures [36].

The goals of treating periprosthetic hip infections are the control of infection and the restoration of function of the affected limb. However, both of these goals have not always been attainable. The extent of infection and the interval for which it has been present play a role in the choice of the revision procedure and the chances for successful treatment after revision [31]. In chronic infections, the gold standard has been a two-stage procedure whereby the primary implants are removed, a radical débridement is performed, and an interim spacer is inserted. Then a second procedure is planned once control of infection is achieved with local and systemic antibiotics [35, 36]. In acute infections, aggressive open débridement with or without exchange of mobile parts (femoral heads and acetabular inserts) and retention of the infected implant has been advocated for early or late infections with a short duration of symptoms, stable components, no evidence of immunosuppression, and overlying soft tissue and skin of good condition [12, 42]. The aim of rapid intervention with thorough open débridement is to prevent the production of any biofilm by the infecting organism, paramount for a successful outcome [31]. Difficulties with this approach, however, include complete débridement of all dead or nonviable tissues and determination of the time of onset of infection and the point beyond which it is no longer reasonable to retain the implant. Davis [12] suggested up to 2 weeks for early infection and up to 72 hours for acute hematogenous, late infection, whereas Zimmerli et al. [42] recommended a period of 3 weeks for both early and late infections. Long-term suppressive antibiotics are indicated when an operation is refused by the patient or is believed to be associated with an unacceptable risk in medically unfit patients, and salvage procedures such as Girdlestone arthroplasty and arthrodesis are considered in life threatening or intractable hip infections and when patient or limb viability is at risk [36].

Open débridement with exchange of modular components has remained a relatively low-morbidity option in acutely infected THAs, allowing less complex surgery and lower cost in comparison to a two-staged procedure. However, infection control rates have varied from 20% to 87% due to the lack of consistent treatment protocols and variations in surgical techniques [3, 4, 8, 11, 32, 34, 39]. Additionally, control of infection with open débridement has been linked to various factors, including host comorbidities, virulence of the infecting organism, timing of the infection in relation to the original surgery, and duration of symptoms before surgery [26].

To supplement the current literature, we determined (1) the rate at which aggressive early débridement with exchange of modular components would control acutely infected cemented THAs, (2) the most common microorganisms responsible for both infections and reinfections, and (3) the final treatment modality resulting in infection control for each patient at latest followup.

Patients and Methods

We performed a retrospective cohort analysis of a prospectively compiled register of all 26 patients with acutely infected cemented THAs (16 primaries, 10 revisions) that occurred within 6 weeks of the index operation (18 patients) or of hematogenous spread (eight patients) between 1999 and 2006. Participants included 10 men and 16 women with a mean age of 69.5 years (range, 51–86 years) (Table 1). The right hip was affected in 17 patients and the left hip in nine. No bilateral infections were included in our study. The primary operation was performed in our institution in 10 patients and elsewhere in 16. The majority of patients had osteoarthritis as the underlying pathology for their primary THA. Acute hematogenous infections were characterized by an acute presentation of symptoms in a previously well-functioning hip arthroplasty. In this group of patients, a full workup to establish the source of infection was undertaken, including a comprehensive history of recent systemic infections or invasive procedures causing bacteremic seeding of the hip, and investigations performed included a throat swab, chest radiograph, and urine, stool, and blood cultures. A source of infection was identified in all patients (one upper respiratory tract infection, one lower respiratory tract infection, six urinary tract infections) and the bacteria isolated in each case was the same bacteria as cultured from the prosthetic joint. In the revision THA group, the original indications for reoperation after primary procedures were aseptic loosening and wear. No patient had a previous history of infection of the affected hip, and none had prosthetic loosening or malalignment at the time of presentation. The mean time between onset of hip symptoms and débridement was 20 days (range, 1–41 days). Comorbidities were assessed according to the American Society of Anesthesiologists (ASA) grading system [25]; three patients were Grade 1, 11 Grade 2, and 12 Grade 3. Diagnosis of infection and a decision to perform a débridement were based on a high index of suspicion from clinical presentation (pain, fever, swelling, skin redness, discharging sinus) and serologic testing (erythrocyte sedimentation rate [ESR] > 30 mm/hour; C-reactive protein [CRP] > 10 mg/L) [36]. Definitive diagnosis was established from hip fluid and at least five tissue samples sent for culture at the time of débridement but before commencement of any antibiotic treatment [36]. The minimum followup was 5 years (mean, 6.6 years; range, 5–11 years). No patients were lost to followup. No patients were recalled specifically for this study; all data were obtained from medical records and radiographs.

Table 1.

Patient details

Patient	Age (years)	Sex	Original operation	Indication	Cementation (original operation)	Type of infection	Time from operation (days)	Symptom duration (days)	Causing organism	Number of débridements	Control of infection	Final treatment	Followup (months)
1	73	Female	Primary	OA	Yes	P	2		MRSA	1	Yes	Débridement	68
2	65	Male	Primary	AVN	Yes	H		3	S aureus	2	No	Two-stage revision
3	69	Female	Primary	OA	Yes	P	32		S aureus	2	Yes	Débridement	76
4	61	Male	Primary	DDH	Yes	P	30		Streptococcus	1	Yes	Débridement	125
5	71	Male	Primary	OA	Yes	H		1	Pseudomonas	2	Yes	Débridement	80
6	72	Female	Primary	OA	Yes	P	31		Propionibacterium spp	1	Yes	Débridement	72
7	66	Female	Primary	DDH	Yes	P	1		Propionibacterium spp	1	Yes	Débridement	69
8	74	Female	Primary	OA	Yes	H		40	Bacteroides	1	Yes	Débridement	88
9	52	Male	Primary	SUFE	Yes	P	31		CNS	2	No	Two-stage revision
10	72	Female	Primary	DDH	Yes	P	2		CNS	1	Yes	Débridement	65
11	69	Female	Primary	OA	Yes	P	29		Streptococcus	1	Yes	Débridement	61
12	64	Female	Primary	OA	Yes	P	1		MRSA	2	Yes	Débridement	82
13	83	Female	Primary	OA	Yes	P	38		Corynebacterium spp	1	Yes	Débridement	80
14	69	Male	Primary	OA	Yes	H		2	Corynebacterium spp	1	Yes	Débridement	112
15	68	Female	Primary	OA	Yes	P	1		Acinetobacter	1	Yes	Débridement	64
16	78	Female	Primary	OA	Yes	H		41	CNS	3	No	Two-stage revision
17	60	Female	Revision	Perthes	Yes	P	22		CNS	1	Yes	Débridement	69
18	67	Female	Revision	OA	Yes	P	2		S aureus	1	Yes	Débridement	62
19	51	Male	Revision	AVN	Yes	H		4	S aureus	1	Yes	Débridement	64
20	86	Male	Revision	OA	Yes	P	28		Pseudomonas	2	No	Two-stage revision
21	56	Female	Revision	DDH	Yes	H		41	CNS	2	Yes	Débridement	133
22	74	Female	Revision	OA	Yes	P	16		CNS	1	No	Antibiotic suppression	84
23	62	Female	Revision	DDH	Yes	P	35		CNS	1	Yes	Débridement	65
24	81	Male	Revision	OA	Yes	H		40	CNS	1	Yes	Débridement	69
25	77	Male	Revision	OA	Yes	P	29		Acinetobacter	1	Yes	Débridement	75
26	75	Male	Revision	OA	Yes	P	19		MRSA	1	No	Two-stage revision

OA = osteoarthritis; AVN = avascular necrosis; DDH = developmental dysplasia of the hip; SUFE = slipped upper femoral epiphysis; P = postoperative; H = hematogenous; CNS = coagulase-negative Staphylococcus; MRSA = methicillin-resistant Staphylococcus aureus.

One of us (FSH) performed all procedures. Patients underwent aggressive débridement including an open complete synovectomy, exchange of all mobile parts, débridement of all aspects of the joint, irrigation with hydrogen peroxide and Betadine^® solutions, and then pulsatile lavage. A drain was inserted in all patients and was removed at 48 hours or earlier if output was less than 50 mL after 24 hours.

Postoperatively, we allowed patients to mobilize with full weightbearing and discharged them when deemed safe. Patients continued antibiotic therapy tailored to the sensitivities of intraoperative cultures for at least 6 weeks until inflammatory markers (CRP, ESR) and the plasma albumin concentration returned to within normal limits. Early conversion to oral antibiotics was dictated by sensitivities and consultation with the microbiology department.

We followed patients at 2 and 6 weeks, 6 months, 1 year, and then on a yearly basis, looking for clinical symptoms and signs of infection, as well as CRP and ESR level testing. Control of infection was defined as absence of clinical, serologic, and radiographic signs of infection and absence of death secondary to infection or treatment during the followup period. We defined failure as any major operation performed after the débridement(s) for control of infection, including a two-stage revision, excision arthroplasty, arthrodesis, and amputation, or the need for long-term antibiotic suppression. We performed up to a maximum of three débridements before proceeding to any further surgical intervention or considering long-term antibiotic suppression for a failure of infection control, taking into consideration patients’ comorbidities and risks for surgery and their preference for choice of treatment. We considered reinfection to be an infection with the same or another organism and treated it aggressively with additional débridement, either alone or with removal of the implants and a staged revision.

We obtained plain radiographs, including an AP pelvis and lateral of both hips, at every followup appointment. Two of the authors (MS, SP) assessed stem position, radiolucencies, and osteolysis. The stem angle was classified as neutral, varus, or valgus. A stem angle was considered neutral if its axis was within 2° of the femoral shaft axis. Femoral and acetabular radiolucencies were classified according to the zones of Gruen et al. [17] and DeLee and Charnley [13], respectively. Loosening was diagnosed if the radiolucent zone around one or both components was 2 mm or more in width and a patient had symptoms on weightbearing and motion that were relieved by rest [18]. Osteolytic lesions were documented and classified on the basis of their size (linear or expansile) and their location according to previously published criteria by Zicat et al. [41]. Substantial interobserver variability can be expected using these criteria [21, 29].

Results

At latest followup, 20 of the 26 patients (77%) achieved control of their infections and returned to their expected functional level with no evidence of recurrence or loosening, wearing away, or malpositioning on followup radiographs. Eight patients had multiple washouts and the infection was controlled in four at latest followup. Five patients underwent a two-stage revision due to reinfection and one patient was placed on long-term antibiotic suppression.

Microbiology confirmed bacterial infection in all patients, with the most commonly isolated organism being coagulase-negative Staphylococcus (eight patients, 31%), followed by S aureus (seven patients, 27%), of which three were methicillin resistant (Table 1). Other microorganisms isolated included Group C Streptococcus, Corynebacterium spp, Propionibacterium spp, Acinetobacter, Pseudomonas aeruginosa, and Bacteroides fragilis (Fig. 1). Most common reinfections were due to coagulase-negative Staphylococcus and S aureus (Fig. 2).

Fig. 1.

A pie chart shows the different microorganisms grown from intraoperative wound swabs and tissue biopsies. MRSA = methicillin-resistant S aureus.

Fig. 2.

A graph shows the microorganisms responsible for reinfections after débridement. CNS = coagulase-negative Staphylococcus.

One of the 10 patients treated within 5 days of symptom onset developed a reinfection. This was in contrast with the remaining 16 patients who received treatment beyond 5 days of onset of symptoms, of whom five had reinfections (Fig. 3). No other variables seemed to influence infection control (Table 1).

Fig. 3.

A graph shows reinfections according to duration of symptoms before the débridement procedure.

Discussion

Management of periprosthetic hip infections remains challenging and costly. A two-stage revision procedure provides the most definitive treatment but with substantial morbidity, taking into consideration two major operations and the overall length of time necessary for intravenous antibiotic treatment [11, 35, 36]. In the acutely infected THA, both patients and surgeons often view open débridement with or without exchange of the modular components as an acceptable alternative. However, the rates for infection control vary in the literature due to the differences in management protocols [3, 4, 32]. Therefore, we (1) determined the rate at which aggressive early débridement with exchange of modular components would control acutely infected THAs, (2) studied the most common microorganisms responsible for both infections and reinfections, and (3) reported the final treatment modality resulting in infection control for each patient at latest followup.

Our study is associated with some limitations. First, infection control after hip arthroplasties can be affected by a number of risk factors, such as age, sex, time from operation, duration of symptoms, patient comorbidities, and the pathogen causing the infection [14, 36, 39]. Because of the small number of patients and observational nature of this study, we were unable to perform a multivariate analysis to further investigate the effect of those risk factors on infection control outcome. Second, the study population was heterogeneous in relation to the type of original operation (primary versus revision) and type of infection (postoperative versus hematogenous).

Our results for infection control (77%) are consistent with those previously reported in the literature (Table 2). Duration from symptom onset has similarly been reported as the most important prognostic factor for infection control in a number of studies [3, 6, 17, 19]. For instance, Crockarell et al. [11] suggested débridement and retention of prosthesis have been successful when treatment was commenced within 6 days of symptom onset. Similarly, Tattevin et al. [37] recommended an interval of 2 to 5 days between the onset of symptoms and surgical treatment. Brandt et al. [7], in another series, suggested a 2-day cutoff for intervention after THA or TKA presenting with acute periprosthetic infection. Meehan et al. [30] and Marculescu et al. [27] suggested an elevated risk of failure if no treatment was commenced within 4 and 8 days of symptoms onset, respectively. It is of note, though, the inclusion and exclusion criteria, as well as management protocols, varied among those studies, occasionally including all four types of periprosthetic infections rather than acute infections alone. Additionally, some of the studies did not differentiate between hips and knees when reporting their results. This resulted in a wide range of infection control rates reported in those studies (Table 2).

Table 2.

Previous studies reporting prosthesis retention after irrigation and débridement

Study	Infection site	Number of patients	Exchange of mobile parts	Retention rate (%)	Followup (years)
Aboltins et al. [1] (2010)	Hip/knee	17	Yes	88.2	2.3
Aboltins et al. [2] (2007)	Hip/knee	20	Partly	90	2.7
Azzam et al. [5] (2010)	Hip/knee	104	29%	44	5.7
Berdal et al. [6] (2005)	Hip/knee	18	Yes	94.5	1.8
Cobo et al. [10] (2010)	Hip/knee	103	Yes	54.3	2.4
Crockarell et al. [11] (1998)	Hip	42	No	26	6.3
Estes et al. [16] (2010)	Hip/knee	20	Yes	90	3.5
Klouche et al. [20] (2011)	Hip	12	Partly	75	3.3
Krasin et al. [23] (2001)	Hip	7	No	71	2.5
Marculescu et al. [27] (2006)	Hip/knee	99	48%	60	2
Martínez-Pastor et al. [28] (2009)	Hip/knee	47	Yes	74.5	1.2
Meehan et al. [30] (2003)	Hip/knee	19	26%	89.5	3.9
Tattevin et al. [37] (1999)	Hip/knee	34	No	38.2	1.6
Tintle et al. [38] (2009)	Hip/knee	8	Yes	100	3.1
Tsukayama et al. [39] (1996)	Hip	41	Yes	68	6.8
Van Kleunen et al. [40] (2010)	Hip/knee	18	72%	72.2	2.6
Zimmerli et al. [43] (1998)	Hip/knee	8	No	100	2.9
Current study	Hip	26	Yes	77	6.6

Identification of the causative organism(s) is an important factor in confirming the diagnosis of infection and dictating perioperative treatment with appropriate antibiotics. However, we believe an early aggressive débridement based on high index of clinical suspicion takes priority over performing preoperative diagnostic aspiration and biopsies and awaiting the results because this not only delays surgical intervention but also carries variable sensitivity and specificity rates for diagnosing infection (0.50–0.93 and 0.82–0.97, respectively [22, 24, 33]). Furthermore, despite the common assumption that gram-negative bacteria are more difficult to treat [36], we only had one case of reinfection with Pseudomonas, suggesting the results of preoperative samples should not delay early surgery for the sake of identifying the infecting organism. Similarly, two of three cases of methicillin-resistant Staphylococcus infections responded to early treatment, with the infection being controlled at 68 and 82 months’ followup. Four patients who underwent two débridements achieved infection control at latest followup, which may provide a less traumatic alternative to the patient than a two-stage exchange.

In conclusion, our data are in agreement with current evidence in the literature and suggest there may be a role for aggressive early débridement in controlling acute postoperative and hematogenous infections after cemented THA.