Middle-term two-stage treatment results of fistulous and non-fistulous form of chronic hip periprosthetic joint infection

全文:

INTRODUCTION

Large joint replacement is a surgery routinely performed worldwide [1, 2]. The number of total hip replacements (THRs) increases annually, as does the incidence of associated complications [3, 4]. Periprosthetic infections (PPIs) are especially concerning. The incidence of PPIs is 1–3% after primary surgeries and increases to 10–15% after revision surgeries [5, 6].

PPIs pose a significant burden to patients and healthcare systems. According to studies, the incidence of recurrent PPIs ranges from 12% to 22% [7, 8], with a one-year mortality rate of up to 5.5% [9]. The costs of PPI therapy in the USA are expected to reach $1.8 billion by 2030 [4].

Two-stage revision arthroplasty is the most prevalent treatment method for patients with chronic PPIs [10], with a particular emphasis on surgical debridement and antibiotic spacer placement. The efficacy of this stage determines the success of revision arthroplasty and the possibility of infection reversal at the end of treatment [7, 10].

Recent studies assess various risk factors for recurrent PPIs, including age, comorbidities, pathogens, and spacer type [5, 8, 11]. Fistulous tract is one of the most important factors. According to the 2018 International Consensus Meeting on Musculoskeletal Infection, fistulous tract is a major PPI criterion [12]; however, its role in a potential recurrence remains unclear. This determined the objective of our study.

The study is aimed to compare the physical status of patients, infection cause, efficacy of chronic periprosthetic hip joint infection therapy, and treatment outcomes depending on the presence of a fistulous tract.

MATERIALS AND METHODS

Study design

A retrospective, single-center cohort study was performed using medical records and phone surveys of 316 patients with hip PPIs who underwent primary hip replacement (Fig. 1). The first stage of treatment was performed in the septic surgery department, and the second stage in specialty orthopedic surgery departments between 2014 and 2018.

 

Fig. 1. Patient selection flow chart

Note. ППИ — periprosthetic joint infection, ТБС — hip joint.

 

Eligibility criteria

Inclusion criteria: newly diagnosed chronic hip PPI; surgical debridement with antibiotic spacer placement.

Non-inclusion criteria: previous revision hip surgeries decompensated concomitant pathology, systemic inflammatory response, and sepsis.

Intervention

The surgery involved radical surgical debridement of the joint cavity: necrectomy and removal of all implant components and bone cement (if any). An antibiotic spacer was placed in the joint cavity, with a Redon drain. The layered closure technique was used to completely close the wound.

Study setting

A total of 218 patients who met the criteria listed above were selected for further assessment. Chronic PPIs were confirmed using the 2018 International Consensus Meeting criteria [12]. The study participants were divided into two groups depending on the presence of a fistulous tract: Group 1 included patients with PPI without fistulas, and Group 2 included patients with PPI with fistulas. The following parameters were assessed in the groups: length of hospital stay, duration of surgical debridement, blood loss, PPI pathogens, comorbidity index [13], infection type by Zimmerli [14], PPI treatment efficacy, treatment outcome, and type of surgery following the index surgery.

PPI causes were assessed based on microbiological testing of tissue biopsies and periprosthetic fluid samples following ultrasonic treatment of removed implants. The type of infection was determined using the classification proposed by Zimmerli. Infections were classified as early (<3 months), delayed (3–12 months), or late (>12 months) based on the time of onset.

Comorbidities were assessed based on the prevalence of various medical conditions. The total comorbidity score was derived by summarizing individual conditions and their severity [13]. The comorbidity index was low (0–7 points), moderate (8–12 points), or high (>12 points), respectively.

A group (n = 202) for assessing treatment efficacy and further surgery depending on the outcome was formed based on medical records, local registry of hip PPIs, and phone survey.

The mean duration of follow-up was 26.2 months (13–38 months). The efficacy of chronic infection treatment was assessed after each surgical stage. Eradication criteria included the absence of clinical and laboratory signs of infection, as well as the absence of a documented recurrence between treatment stages, according to the Consensus Meeting criteria.

Statistical analysis

Data were processed with StatSoft Statistica 10. The frequencies of qualitative parameters (sex, PPI type, treatment efficacy) were compared using the Pearson’s chi-squared test. Numerical values were presented as the median (Me) and interquartile range (Q1–Q3; 25–75%). Differences in quantitative parameters (age, length of hospital stay, duration of surgery, total comorbidity score) were assessed using the Mann–Whitney test. Differences were considered significant at p < 0.05.

Ethics approval

The study followed the ethical standards of the World Medical Association’s Declaration of Helsinki, as revised by the Ministry of Health of the Russian Federation. The patients provided informed consent for participation in the study and the publication of anonymized study findings.

RESULTS

The mean age of patients in the general cohort was 61 years (Table 1), with 56% being female. Patients with a fistulous tract were significantly younger (p = 0.006). Late infections were the most common infections detected in the study (n = 218). According to intergroup analysis, Group 2 had a significantly higher incidence of PPI 3–12 months following hip replacement (p = 0.002).

 

Table 1. Characteristics of the patients

Parameter	Total (n=218)	Without fistulas (n=80)	With fistulas (n=138)	p-value
Male, n (%)	96 (44)	35 (44)	61 (44)	>0,05
Female, n (%)	122 (56)	45 (56)	77 (56)
Age, years (mean)	50–69 (61)	53–73 (63)	45–67 (58)	0,006
PPI type, n (%)
Early	51 (23)	14 (18)	37 (27)	>0,05
Delayed	49 (22)	13 (16)	36 (26)	0,02
Late	118 (54)	53 (66)	65 (47)	>0,05

Note: PPI, periprosthetic infection. Significant parameters are given in bold.

 

The comorbidity index was 8 points in both groups. Patients in Group 1 required a significantly longer hospital stay (mean: 26 days; p = 0.03). The duration of surgery and blood loss were insignificantly greater in Group 2 compared to Group 1 (Table 2).

 

Table 2. Studied indicators in comparison groups

Parameter	Without fistulas (n=80)	With fistulas (n=138)	p-value
Comorbidity index	8 (6–11)	8 (5–11)	>0,05
Length of hospital stay, days	26 (21–35)	23 (20–29)	0,03
Duration of surgery, minutes	180 (150–208)	190 (165–225)	>0,05
Intraoperative blood loss, mL	700 (500–1250)	800 (600–1000)	>0,05

Note: Significant parameters are given in bold.

 

The analysis of PPI causes found that staphylococci were the most common pathogens (51%; Table 3). S. epidermidis was the most prevalent pathogen in Group 1 (p < 0.001), and S. aureus in Group 2 (p < 0.001). Moreover, rare pathogens from the “Other” category were significantly more common in Group 2 (p = 0.04). Other than that, the etiology was comparable in the two groups.

 

Table 3. The periprosthetic joint infection pathogen structure in the study groups, n (%)

Pathogen	Without fistulas	With fistulas	p-value
Staphylococcus epidermidis	43 (34)	31 (18)	<0,001
Staphylococcus aureus	16 (17)	72 (43)	<0,001
Enterococcus sp.	10 (10)	9 (5)	>0,05
Streptococcus sp.	8 (8)	8 (5)	>0,05
NFGNB	6 (6)	11 (7)	>0,05
Enterobacteriaceae	3 (3)	8 (5)	>0,05
CoNS	3 (3)	9 (5)	>0,05
Corynebacterium sp.	3 (3)	2 (1)	>0,05
Candida sp.	2 (2)	2 (1)	>0,05
Propionibacterium sp.	1 (1)	4 (2)	>0,05
Other	1 (1)	11 (7)	0,04
Total	96 (100)	167 (100)

Note: NFGNB, non-fermenting Gram-negative bacteria: P. aeruginosa, Acinetobacter spp.; Enterobacteriaceae, including Escherichia coli, Klebsiella pneumoniae, Enterobacter cloacae; CoNS, coagulase-negative staphylococci (except for S. epidermidis); other, Actinomyces spp., Eggerthella spp., Alcaligenes spp., Bacteroides spp., etc. Significant parameters are given in bold.

 

The prevalence of culture-negative PPIs was 5% in both groups (Fig. 2).

 

Fig. 2. Distribution of periprosthetic infection types in the study groups, p > 0.05

Note: PPI, periprosthetic infection

 

The prevalence of polybacterial PPIs was comparable; however, microbial associations with Gram-negative pathogens were 1.8 times more common in Group 2 (p > 0.05) (Table 4).

 

Table 4. Incidence of mono- and polybacterial periprosthetic infections, n (%)

Parameter	Without fistulas	With fistulas	p-value
Monobacterial PPI
Gr(+)	55 (87)	92 (91)	>0,05
Gr(-)	6 (10)	8 (7)	>0,05
Fungi	2 (3)	1 (2)	>0,05
Polybacterial PPI
Without Gr(-)	10 (77)	17 (56)	>0,05
With Gr(-)	3 (23)	13 (43)

Note (here and in Table 5): PPI, periprosthetic infection; Gr(+), Gram-positive pathogens; Gr(-), Gram-negative pathogens.

 

The efficacy of PPI treatment after the first stage in the general cohort was 78% (Fig. 3). The efficacy of PPI treatment after the second stage was 99% (Fig. 4).

 

Fig. 3. Recurrence rate after the first and second surgical debridement

Note: *, p < 0.05.

 

Fig. 4. Treatment outcomes

Note: PPI, periprosthetic infection; RHA, revision hip arthroplasty; SE, spacer exchange; RWD, radical wound debridement; SSE, second spacer exchange; MP, myoplasty.

 

Recurrent chronic infections after surgical debridement were insignificantly more common in Group 2 (p > 0.05). Further surgical treatment of recurrent PPIs significantly reduced treatment efficacy: from 78% for the first debridement to 61% for the repeated debridement (p = 0.02). The presence of a fistulous tract significantly affected the efficacy of recurrent infection treatment (p = 0.048).

Various surgical techniques were used to treat recurrent PPIs, including radical wound debridement (RWD), antibiotic spacer exchange, and excision arthroplasty in combination with myoplasty using a vastus lateralis muscle flap (Fig. 4). Despite the small sample size, the outcomes of various surgical treatments in the first recurrent PPI with fistulas are of interest. RWD with preserved antibiotic spacer resulted in PPI eradication only in 33% of cases, compared to spacer exchange (64%) and myoplasty (75%) (p > 0.05).

The follow-up of patients with a second PPI recurrence revealed differences in treatment outcomes (p > 0.05). In Group 1, PPIs resolved in 4 out of 5 cases. In Group 2, there was one fatal outcome prior to treatment. The outcomes in four patients with second spacer exchange were unfavorable: one fatal outcome and two cases of third recurrence (Fig. 3).

The cause of infection had no significant impact on the risk of first recurrence. Recurrent PPIs with fistulas caused by Gram-negative bacteria or microbial associations were 1.8 times and 3 times more common, respectively (p > 0.05) (Table 5).

 

Table 5. Recurrence rate in patients with mono- and polybacterial periprosthetic infections in the study groups, n/N (%)

PPI cases	Recurrence		p-value
	Without fistulas	With fistulas
Mono- Gr(+)	10/52 (20)	21/85 (25)	>0,05
Mono- Gr(-)	1/6 (16)	2/7 (29)	>0,05
Poly- without Gr(-)	1/9 (11)	6/22 (27)	>0,05
Poly- with Gr(-)	0/3 (0)	1/6 (17)	>0,05
No growth	0/4 (0)	1/6 (17)	>0,05
Fungi	1/2 (50)	0/2 (0)	>0,05
Total	13/74 (17,6)	31/128 (24,2)	>0,05

 

DISCUSSION

Chronic periprosthetic hip joint infections continue to be a challenge in orthopedic surgery. The efficacy of infection treatment depends on the extent of debridement and adequate antibiotic therapy. In our study, the efficacy of surgical debridement stage in the study cohort with chronic PPIs was 78%, which is consistent with published data [7–9]. However, multiple factors can affect the outcome. In particular, the role of fistulous tract in infections remains debatable. Researchers agree that fistulas are caused by long-term inflammation in the artificial joint area. Chronic infection creates a tract that connects the implant to the skin surface. This finding is used as a major PPI criterion [12] that requires no further proof.

According to Parvizi et al., fistulous tract formation is associated with smoking, hypothyroidism, hypoalbuminemia, and previous revision surgeries. The presence of a fistulous tract increased the risk of recurrence in two-stage surgery by almost 3 times [15]. Other researchers reported a similar tendency [16]. In our study, there were no significant differences in comorbidities, and previous revision surgeries were a non-inclusion criterion.

It is worth noting the difference in the mean age of patients. The mean age in the study cohort (n = 218) was 61 years, which corresponded to the mean age of patients after hip replacement [2]. However, patients in Group 2 were significantly younger than patients in Group 1, with the mean age of 58 years and 63 years, respectively (p = 0.006). Fistulous tracts most commonly formed 3–12 months after surgery, which is likely due to increased responsiveness in this group of patients.

The analysis of PPI causes found that staphylococci were the most common pathogens, as expected. This is consistent with worldwide tendencies [17]. S. epidermidis was more frequently isolated in Group 1, and S. aureus in Group 2. Moreover, rare pathogens (Actinomyces spp., Eggerthella spp., Alcaligenes spp., Bacteroides spp.) were more frequently observed in Group 2, which may be due to bacteria infiltrating the prosthetic joint area from the skin via the fistulous tract.

Published data suggest that the probability of isolating microbial associations is higher in patients with fistulas [17]. In our study, there were no significant differences in the incidence of mono- and polybacterial PPIs. The incidence of polybacterial PPIs in Groups 1 and 2 was 16% and 22%, respectively. Microbial associations with Gram-negative bacteria were almost 2 times more common in Group 2 (43% vs 23%; p > 0.05). The presence of a fistulous tract in patients with polybacterial PPIs increased the risk of an unfavorable outcome from 8% to 25% (p > 0.05). Other researches have shown that microbial associations have a detrimental impact on treatment outcomes, which supports our findings [18, 19].

In our study, Group 1 required a longer hospital stay than Group 2: 26 days and 23 days, respectively (p = 0.03). This was likely due to the method of PPI diagnosis [12]. In patients without a fistulous tract, a hip joint puncture was performed; the time spent waiting for the results increased the length of hospital stay by an average of 3 days.

The average duration of surgery and blood loss during surgical debridement for hip PPIs were comparable in both groups, amounting to approximately 3 hours and 800 mL, respectively. The presence of a fistulous tract had no significant impact on the debridement stage, and there were no difficulties with wound closure because of available soft tissues in the hip joint area. In comparable knee joint surgeries, fistulous tracts had a significant impact and caused additional difficulties during the procedure [20]. The average intergroup difference in the duration of surgery and blood loss was 40 minutes and 350 mL, respectively.

International studies show that the efficacy of the first stage of PPI treatment varies significantly. For example, a review by Yermakov et al. revealed that the success rate of two-stage PPI treatment ranges between 75% and 100% (mean: 91%) [21]. Palmer et al. reported similar findings, with an eradication rate of 75% to 100% [22].

An intergroup analysis of treatment efficacy in patients with and without fistulas showed no significant differences. During the second stage of treatment, PPIs resolved in 82% of patients in Group 1 and 76% of patients in Group 2. Differences with comparable knee joint surgeries are most likely attributable to anatomical features [20]. In contrast to the knee joint, the hip joint’s abundance of soft tissues may keep pathogens on the skin surface from joining the existing pathogen and exacerbating the infection.

Many researchers agree that the first stage of treatment is essential for the outcome. According to Wichern et al., the 95% remission rate after the first stage of treatment resulted in a comparable rate after the second stage [23]. In our study, 158 patients underwent revision hip arthroplasty during the second stage, with a 1% recurrence rate.

Recurrent PPIs are especially challenging. For example, in a study by Kozaliy et al., the rate of unfavorable outcomes after spacer exchange was 49% [24]. Klemt et al. reported this outcome in 24% of cases; moreover, it significantly exceeded 15% following primary PPI treatment [25]. In our study, recurrent PPIs were significantly more common than primary infections (39% vs 22%; p < 0.05), with fistulas having a sustained negative impact on the outcome.

The analysis of treatment failures with various surgical techniques revealed that repeated surgical debridement is ineffective, with a 66% recurrence rate. Despite the small sample size in our study (n = 6), RWD appears to be an inferior choice when selecting treatment strategy in recurrent infections, which is supported by other studies [26].

CONCLUSION

PPIs with fistulas are more common in younger patients and require less time to confirm the diagnosis. Fistulous tract has no impact on the recurrence rate; however, it impairs treatment outcomes in later recurrences. Repeated surgical debridement in recurrent PPIs appears to be ineffective, which should be confirmed in future studies.

ADDITIONAL INFO

Autor contribution. A.A. Kochish — data collection and analysis, data statistical processing, manuscript writing; S.A. Bozhkova — research concept and design, manuscript writing and editing; V.A. Artyukh — treatment the patients, manuscript editing. All authors confirm that their authorship meets the international ICMJE criteria (all authors have made a significant contribution to the development of the concept, research and preparation of the article, read and approved the final version before publication).

Funding source. The authors state that there is no external funding when conducting the research and preparing the publication.

Competing interests. The authors declare that they have no competing interests.

作者简介

Andrey Kochish

Vreden National Medical Research Center of Traumatology and Orthopedics

编辑信件的主要联系方式.
Email: kochishman@gmail.com
ORCID iD: 0000-0001-8573-1096
SPIN 代码: 3717-1640

MD, Cand. Sci. (Medicine)

俄罗斯联邦, 8 Akademika Baykova str., 195427 St. Petersburg

Svetlana Bozhkova

Vreden National Medical Research Center of Traumatology and Orthopedics

Email: clinpharm-rniito@yandex.ru
ORCID iD: 0000-0002-2083-2424
SPIN 代码: 3086-3694

MD, Dr. Sci. (Medicine), professor

俄罗斯联邦, 8 Akademika Baykova str., 195427 St. Petersburg

Vasily Artyukh

Vreden National Medical Research Center of Traumatology and Orthopedics

Email: artyukhva@mail.ru
ORCID iD: 0000-0002-5087-6081
SPIN 代码: 7412-5114

MD, Dr. Sci. (Medicine)

俄罗斯联邦, 8 Akademika Baykova str., 195427 St. Petersburg

参考

补充文件