Periprosthetic joint infection after hemiarthroplasty for hip fracture is a distinct clinical entity associated with high mortality

Abstract

Aims

Periprosthetic joint infections (PJIs) are devastating complications. Our knowledge of hip fracture-associated hemiarthroplasty PJI (HHA-PJI) is limited compared to elective arthroplasty. The goal of this study was to describe the epidemiology, risk factors, management, and outcomes for HHA-PJI.

Methods

A population-based (465,000 patients) multicentre, retrospective analysis of hip hemiarthroplasty (HHA) between January 2006 to December 2018 was conducted. PJI was defined by international consensus and treatment success as no return to operating room and survival to 90 days after the initial surgical management of the infection. Univariate, survival, and competing risk regression analyses were performed.

Results

In total, 1,852 HHAs were identified (74% female; mean age 84 years (SD 7); 90-day mortality 16.7%). A total of 43 patients (2.3%) developed PJI at aged 77 years (SD 10) (56% female; 90-day-mortality 20.9%; hazard ratio (HR) 1.6; 95% CI 1.1 to 2.3; p = 0.023). The incidence of HHA-PJI was 0.77/100,000 population/year and 193/100,000 HHAs/year. The median time to PJI was 26 days (IQR 20 to 97), with 53% polymicrobial growth and 41% multidrug-resistant organisms (MDRO). Competing risk regression identified younger age (subdistribution hazard ratio (SHR) 0.86; 95% CI 0.8 to 0.92; p < 0.001), chronic kidney disease (SHR 3.41; 95% CI 1.36 to 8.56; p = 0.013), BMI > 35 kg/m² (SHR 6.81; 95% CI 2.25 to 20.65; p < 0.001), perioperative urinary tract infection (SHR 1.89; 95% CI 1.02 to 3.5; p = 0.042), and dementia (SHR 9.4; 95% CI 2.89 to 30.58; p < 0.001) as significant risk factors for developing HHA-PJI. When infection treatment was successful (n = 15, 38%), median survival was 1,632 days (IQR 829 to 2,084), as opposed to 215 days (IQR 20 to 1,245) in those who failed, with a 90-day mortality of 30% (n = 12). There was no significant difference in treatment success between debridement, excision arthroplasty, or revision arthroplasty.

Conclusion

HHA-PJI is an uncommon complication, but is significantly associated with mortality. All currently identified predictors are nonmodifiable. Due to the common polymicrobial and MDRO infections, our standard antibiotic prophylaxis may not be adequate. HHA-PJI is a different disease compared to elective PJI with distinct epidemiology, pathogens, risk factors, and outcomes, which require targeted research specific to this unique population.

Cite this article: Bone Jt Open 2025;6(11):1409–1415.

Introduction

Periprosthetic joint infections (PJIs) are a devastating complication following hip arthroplasty. Although hip arthroplasty is primarily used in the elective setting, it plays a key role in the management of femoral neck fractures.^1,2 And while contention still remains as to the exact role for hemiarthroplasty (HHA) and total hip arthroplasty (THA) in the management of geriatric hip fractures (GHF), hemiarthroplasty remains a mainstay of treatment for the lower demand, frailer patients with less than five years life expectancy.^1,3,4 Major hip fracture registries around the globe indicate that hemiarthroplasty constitutes 71% to 94% of arthroplasty in the GHF cohort.⁵ Published prospective and retrospective studies suggest approximately 2% to 3% of HHAs develop a PJI, which is similar to German Joint Registry data, while many other international registries either do not report data specific to hemiarthroplasty, or do not completely capture patients who do not undergo any exchange of componentry and therefore under-report the true PJI rate.^1,2,6,7 As the number of GHFs increase, driven in part by an increasing life expectancy in developing countries, so too will prevalence of HHA-PJIs.^1,8

The majority of literature on PJI, including epidemiology and treatment, focuses on the elective setting.⁹ Hip HHA-PJI and elective (e)THA PJI are often seen as the same, but important clinical and epidemiological differences between eTHA patients and HHA for GHF exist. Elective THA patients are often younger, with fewer medical comorbidities than their GHF HHA counterparts, and due to the nature of unplanned surgery, HHA patients cannot be optimized preoperatively in the same way, nor can surgeons select patients in the same way they can for eTHA.^10-12

Current consensus for primary elective THA PJI recommends the use of debridement, antibiotics, and implant retention (DAIR) and/or revision arthroplasty (rTHA) in one or two stages, depending on the timing and nature of the infection and patient factors.^9,13 Yet, despite the crucial place for hemiarthroplasty in modern orthopaedics, there is no consensus or convincing evidence on the management of HHA-PJI. This leaves a myriad potential treatments with a lack of epidemiological data to support clinicians in decision making. The goal of this study was to describe the contemporary population-based epidemiology, risk factors, management, and outcomes for HHA-PJI.

Methods

Study setting

The study was conducted at two university affiliated hospitals in New South Wales, Australia: The John Hunter Hospital (JHH), a level 1 trauma and tertiary referral centre which manages more than 450 GHFs each year, of which approximately 105 are HHA; and The Maitland Hospital (TMH), a metropolitan university affiliated hospital which manages 125 GHFs each year (40 HHAs per year). These two hospitals service a population of more than 465,000 people, and are the only two hospitals that perform emergency hip fracture surgery in the region; all joint infection complications are managed by one infectious diseases service. Patients who underwent a HHA between January 2006 and December 2018 were identified through a prospectively collected institutional (JHH) database of long bone fractures, the Australia New Zealand Hip Fracture Registry (ANZHFR),¹⁴ and cross-referenced with operative databases at both sites . The minimum follow-up was four years from index surgery. To prevent missing patients due to coding errors, all patients who underwent a HHA in the period were reviewed for any evidence of a PJI in either the index admission or subsequent admissions/presentations to any public emergency department/outpatient clinics in New South Wales. Ethical approval was obtained from Hunter New England Health Ethics Committee (Ethics number HNEHREC: AU202003-09 approved on 30 March 2020).

Population and patient characteristics

Patients included were aged 55 years or older who sustained a low-energy hip fracture and underwent a hip hemiarthroplasty at either institution. No limitation was placed on the prosthesis used, nor the approach. All HHA patients were managed in a multidisciplinary team including orthopaedic surgeons and consultant physicians (general or orthogeriatric). All patients received appropriate antibiotic prophylaxis, typically with 2 grams of cefazolin on induction.

PJI was defined by the major criteria of the 2018 Musculoskeletal Infection Society (MSIS) definition.¹⁵ This was selected as the presence of a discharging sinus, or two positive deep tissue cultures, is widely accepted as constituting a deep infection even though neither the major nor minor criteria of the MSIS definition have been validated in HHA-PJI.

Between January 2006 and December 2018, 1,852 HHAs were performed on 1,713 patients across the two sites. The mean age was 84 years (SD 7), and 74% of the cohort were female. The median American Society of Anesthesiologists (ASA) grade¹⁶ was 3 (IQR 3 to 4) and Charlson Comorbidity Index (CCI)¹⁷ was 2 (IQR 0 to 2). The 90-day mortality rate was 16.7%. The median time from presentation with GHF to index surgery was 33 hours (IQR 19 to 72), with no difference between the two hospitals (JHH 26 hours (19 to 53) vs TMH 88 hours (26 to 125); p = 0.064). At the time of index surgery, 251 patients (14%) had a urinary tract infection (UTI), compared to 15 (0.8%) who had a diagnosis of pneumonia.

Data collection

The medical records of all patients who underwent a hemiarthroplasty within the study period were analyzed for any re-presentations or reoperations for, or suspected HHA-PJI. Data were collected relating to patient demographic characteristics at both the index operation and re-presentations. Collected data included the ASA grade, CCI, presence of any formal diagnosis of ischaemic heart disease (IHD), chronic kidney disease (CKD) stage 3 or 4, diabetes mellitus (DM), BMI of 35 kg/m² and above, dementia, chronic obstructive pulmonary disease (COPD), perioperative antiplatelet or anticoagulant, and usual residence (residential aged care facility (RACF) or home), as well any diagnosis of perioperative UTI. Patients who were identified as having a PJI also had relevant data from their index operation including prosthesis details, as well as data relating to their PJI, and subsequent treatment extracted from the medical records for analysis. Mortality data were extracted from the New South Wales (NSW) Registry of Births, Deaths and Marriages.¹⁸

Outcomes

The primary outcome of this study was a diagnosis of PJI after hip hemiarthroplasty. The secondary outcome was treatment success which was defined as 90-day survival after the initial surgical treatment for HHA-PJI and no unplanned returns to the operating room for infection within 90 days.

Statistical analysis

Continuous data were assessed for distribution and presented as mean and SD or median and IQR. Categorical data are presented as a count and percentage. Risk factors for developing HHA-PJI were assessed using univariate logistic regression. Independent variables with a p-value < 0.4 were entered into a multivariate competing risk regression model, due to the high mortality associated with geriatric hip fracture. Variables were excluded based upon the Wald chi-squared test. Death was seen as a competing event according to the Fine and Gray model. No events were censored with the assumption that this population would not relocate from the region, and hence both death and PJI would be detected. Sub-distribution hazard ratios (SHRs; the probability of an event occurring over time) and 95% CIs are presented for interpretation. The primary outcome of treatment success was assessed using a similar competing risk-regression. Statistical analyses were programmed using Stata v. 13.0 (StataCorp, USA). Statistical significance was set at p < 0.05.

Results

Presentation with PJI

Overall, 43 patients (2.3%) developed a PJI at a median time to diagnosis of 26 days (IQR 20 to 97). The incidence per year was 0.77 HHA-PJI/100,000 population, or when considering the incidence per hemiarthroplasty, 193 PJI/100,000 HHAs. The mean age in the HHA-PJI group was 79 years (SD 10), 77% were female, and the 90-day mortality rate was 20.9% (HR 1.6; 95% CI 1.1 to 2.3; p = 0.023). The 12-month mortality rate was 35% in the HHA-PJI group compared to 28% in the no PJI group (HR 1.28; 95% CI 0.76 to 2.14; p = 0.348). There was no difference between HHA-PJI and no-PJI groups in regard to overall mortality (61% vs 76%; HR 0.74; 95% CI 0.5 to 1.09; p = 0.124) (Figure 1).

Fig. 1.

Survival. PJI, periprosthetic joint infection.

Signs and symptoms on diagnosis of HHA-PJI, as well as laboratory results are detailed in Table I. The most common femoral component in the HHA-PJI group was a cemented Exeter femoral component (Stryker, USA) in 21 cases (49%) and an uncemented Austin-Moore femoral component in 18 cases (42%) (Stryker).

Table I.

Presentation.

Variable	Data
Symptom, n (%)
Hip/groin pain	33 (77)
Wound erythema	27 (63)
Wound discharge/ooze	25 (58)
Sinus	16 (37)
Fever > 38.5°C	10 (23)
Systolic blood pressure < 90 mmHg	6 (14)
Delirium	16 (37)
Shock	7 (16)
Dislocation	6 (14)
Investigations on admission
Mean WCC (SD)	11.7 (4.1)
Mean neutrophils (SD)	9.5 (3.9)
Median CRP, mg/l (IQR)	79 (142 to 186)
Median ESR, mm/hr (IQR)	54 (30 to 73)

WCC, white cell count.

Seven patients (16%) had blood cultures taken at the time of presentation of their PJI, of which six were positive and all bar one (Streptococcus milleri) corresponded to the eventual deep tissue samples from the PJI (blood culture organisms: Enterobacter cloacae (n = two), methicillin-sensitive Staphylococcus aureus (MSSA) (n = two), S. milleri, and Streptococcus agalactiae). Of those who had concomitant infections at the time of index surgery, two patients with pneumonia and 14 with a UTI developed a PJI. However, none of the urine or sputum samples correlated with the causative organism of PJI.

Joint aspiration was performed in 17 (40%) patients before proceeding to the operating room. Bacteria were isolated from 16 (94%) of the aspirates, with 12 of these (71%) matching the eventual deep intraoperative specimens perfectly (MSSA, n = six; methicillin-resistant S. aureus (MRSA), n = one; Actinomyces sp.+ Corynebacterium pseudodiphtheriticum, n = one; Pseudomonas aeruginosa, n = one; Cutibacterium avidum n = one; S. milleri, n = one; and Staphylococcus epidermidis, n = one). The other four positive isolates corresponded to eventual growth, however in the context of polymicrobial PJI.

Multidrug resistant organisms were isolated in 18 cases (41%) (MRSA = seven, coagulase-negative staphylococci (CoNS) = three, vancomycin-resistant Enterococcus (VRE) = four, P. aeruginosa = one, Escherichia coli = one, Morganella morganii = one, Granulicatella adiacens = one), and 23 patients (53%) had polymicrobial growth on intraoperative samples. The most commonly identified group of organisms was gram-negative rods in 21 cases, including P. aeruginosa (n = seven), Enterobacter cloacae (n = three), E. coli (n = three), M. morganii (n = two), Proteus mirabilis (n = two), Serratia marcescens (n = two), and two instances with other bacteria. S. aureus was the isolated organism in 19 cases, followed by coagulase-negative S. aureus (n = 12), MSSA (n = 11), and MRSA (n = 7; Figure 2).

Fig. 2.

Microbiology of hemiarthroplasty periprosthetic joint infection. Cor, Corynebacterium; CoNS, coagulase-negative Staphylococcus species; Fun, fungi; GPC, gram-positive cocci (excluding S. aureus); GPR, gram-positive rod; GNR, gram-negative rod; MRSA, methicillin-resistant S. aureus; MSSA, methicillin-sensitive S. aureus; Spp, Streptococcus species; VRE, vancomycin-resistant Enterococcus.

Risk factors for PJI

Overall, there was a significantly higher prevalence of DM, CKD (class III & IV), BMI > 35, and perioperative UTI in the PJI group (Table II). Younger age at time of index operation was also associated with PJI; of the 15 patients included under the age of 70 years, eight developed a PJI (OR 58.8, 95% CI 20.15 to 170.29, p < 0.001). This PJI subgroup was notably comorbid: 88% female, 38% had a perioperative UTI, 25% had pneumonia, 50% had diabetes with end-organ complications, and 50% had an ASA grade 4. Sex, usually residing in an RACF, ASA grade, and dementia were not significant risk factors on univariate analysis. Multivariate competing risk regression identified younger age, CKD, BMI > 35 kg/m², UTI, and dementia as significant risk factors for developing HHA-PJI (Table III).

Table II.

Prevalence of conditions in the periprosthetic joint infection (PJI) group.

Variable	No PJI (n = 1,809)	Infection (n = 43)	Univariate analysis, OR (95% CI)	p-value*
Female sex, n (%)	1,327 (73)	33 (77)	0.83 (0.41 to 1.71)	0.0619
Mean age, yrs (SD)	84 (7)	79 (10)	0.91 (0.87 to 0.95)	< 0.001
RACF, (n,%)	757 (42)	20 (47)	1.21 (0.66 to 2.22)	0.541
Median CCI (IQR)	2 (0 to 2)	2 (1 to 2)	1.08 (0.85 to 1.39)	0.528
Median ASA grade (IQR)	3 (3 to 3)	3 (3 to 4)	1.16 (0.58 to 2.30)	0.685
DM, n (%)	337 (19)	16 (37)	2.59 (1.38 to 4.86)	<0.001
CKD, n (%)	225 (12)	10 (23)	2.13 (1.04 to 4.39)	0.040
Dementia, n (%)	242 (13)	8 (19)	1.48 (0.68 to 3.23)	0.319
COPD, n (%)	301 (17)	8 (19)	1.15 (0.53 to 2.49)	0.724
Antiplatelet/anticoagulant, n (%)	823 (49)†	24 (56)	1.31 (0.71 to 2.41)	0.382
BMI > 35 kg/m2, n (%)	179 (10)	10 (23)	2.76 (1.34 to 5.69)	0.006
UTI, n (%)	333 (18)	14 (33)	2.14 (1.12 to 4.09)	0.018
Pneumonia, n (%)	103 (6)	3 (7)	1.24 (0.38 to 4.08)	0.720

^*Wald z-test.

^†Total n = 1,678.

ASA, American Society of Anesthesiologists grade; CCF, congestive cardiac failure; CCI, Charlson Comorbidity Index; CKD, chronic kidney disease (stage III/IV); COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus; OR, odds ratio; RACF, residential aged care facility resident; UTI, perioperative urinary tract infection.

Table III.

Risk factors for developing hemiarthroplasty periprosthetic joint infection.

Competing risk regression	SHR	95% CI	p-value*	Coefficient	SE
Age	0.86	(0.80 to 0.92)	< 0.001	-4.30	0.03
CKD	3.41	(1.36 to 8.56)	0.013	2.61	1.60
Dementia	9.40	(2.89 to 30.58)	< 0.001	3.72	5.66
Antiplatelet/anticoagulant	1.63	(0.33 to 3.02)	0.124	1.54	0.51
BMI > 35 kg/m2	6.81	(2.25 to 20.65)	< 0.001	3.39	3.85
UTI	1.89	(1.02 to 3.50)	0.042	2.02	0.59

^*Wald z-test.

CKD, chronic kidney disease (stage III/IV); SE, standard error; SHR, sub-hazard ratio; UTI, perioperative urinary tract infection.

Treatment

Overall, the 43 PJIs were managed with 72 operations in total. Four patients were managed without any operative intervention. There were 43 debridements, 11 excision arthroplasty (Girdlestone), six debridements with exchange of modular components, and six exchange/revision arthroplasties. When considering initial operation, the most common operation was debridement (n = 22), followed by excision arthroplasty (n = eight), debridement with exchange of modular components (n = six), and arthroplasty (n = three).

All patients were started on intravenous antibiotics after the initial operation; flucloxacillin was the most common (ten cases), followed by vancomycin (nine cases), cefazolin (five cases), and piperacillin + tazobactam (three cases). Only 21 cases (49%) were placed on an initial antibiotic that correlated with the sensitivities for the later-isolated organism. The mean duration of intravenous antibiotics was 4.4 (SD 2.5) weeks followed by an average of 5.9 weeks (SD 3.2) of oral antibiotics. Overall, ten patients died while on IV antibiotics. Lifelong suppression was selected in nine patients, of which eight had successful treatment.

The most successful initial operation was revision arthroplasty (66%), followed by debridement with exchange of modular components (50%), debridement alone (32%), and excision arthroplasty (25%). And when considering each individual surgery, overall treatment success was low: 50% for debridement with exchange of modular components, 45% for excision arthroplasty, 33% for exchange/revision arthroplasty, and 26% for debridement alone. In those with successful treatment, median survival was 1,632 days (IQR 829 to 2,084) compared with 215 days (IQR 20 to 1,245) (HR 0.47; 95% CI 0.21 to 1.04; p = 0.063). There was no difference, however, between treatment success and the adequacy of the initial antibiotic selection (43% vs 22%; OR 1.95, 95% CI 0.5 to 7.5, p = 0.33).

Discussion

Hip hemiarthroplasty-associated PJI following hip fractures is a different clinical condition to PJI following elective total hip arthroplasty. This study presents one of the largest population-based samples of HHA-PJI. The incidence of 2.3% of HHA-PJI in the GHF cohort is similar to previously published data, and our study also identified a previously unreported population-based incidence of HHA-PJI of 193/100,000 HHA per year. It should be noted that the true incidence of HHA-PJI is likely to be far greater than 2.3%, as many of these frail patients are likely to die before representation or diagnosis of PJI. This cohort was a representative sample of a typical GHF population with female dominance, mean age of 84 years with multiple comorbidities, high anaesthetic risk, and 90-day mortality. All patients included received orthogeriatric input at the time of index surgery and were operated acutely within the internationally accepted timeframe for hip fracture surgery.¹⁹

As opposed to eTHA /total knee arthroplasty (TKA)-PJI, HHA-PJI appears to be almost exclusively an acute/hyperacute infection. Median time from index HHA to diagnosis was just 26 days, making intraoperative or early postoperative inoculation the most likely portal of entry rather than haematogenous spread. As has been seen with acute eTHA/TKA-PJI, gram-negative species are common. S. aureus species (MSSA + MRSA) were the most commonly isolated species; however, gram-negative rods were the most common group isolated (including P. aeruginosa, E. cloacae, E. coli, M. morganii). Unlike the eTHA/TKA cohort, polymicrobial infection or multidrug-resistant organisms are far more common in the HHA-PJI cohort; polymicrobial infection has been shown to be 10% in an eTHA/TKA cohort (PIANO trial) compared with 53% in this study and a MDRO rate of 41% in the HHA-PJI group.²⁰

Even the clinical presentation of HHA-PJI is different to eTHA/TKA-PJI. The PIANO trial²⁰ showed acute eTHA/TKA-PJI to be associated with a higher CRP and WCC (230 mg/dl vs 79 mg/dl, and 13 × 10^9 /l vs 11.7 × 10^9 /l, respectively). Septic shock was seen in 15% of HHA-PJI compared with 7.7%, while a draining sinus was seen in 37% of HHA-PJI compared with just 1.7% of eTHA/TKA-PJI. Interestingly, dislocation was seen in 14% of HHA-PJI, an associated complication of PJI not reported in PIANO. These differences suggest the HHA-PJI cohort may not be able to mount the same clinical response to infection, with higher rates of septic shock despite lower laboratory markers. It is important as clinicians hold a high index of suspicion for PJI in the HHA cohort; septic shock can be difficult to localize, and an ‘oozy wound’ with a modestly elevated CRP two to three weeks postoperatively may both reflect HHA-PJI.

This study used both univariate and multiple competing risk regression (CRR) modelling to identify risk factors for HHA-PJI. Univariate analysis identified diabetes, CKD, BMI > 35 kg/m², UTI at the time of index surgery, and younger age as risk factors for PJI. Given the high mortality rate in the GHF cohort, the authors thought it prudent to treat death as competing risk of developing a PJI and CRR analyses. These identified younger age, CKD, BMI, UTI at time of index surgery, and dementia as risk factors for HHA-PJI. Younger patients in the GHF cohort may be more prone to infection for reasons beyond the scope of this study; however, previous literature suggests that a large proportion of younger hip fracture patients have so called ‘biologically advanced age’ – their bodies are older than their chronological age. This is due to various factors, including a higher rate of complex comorbidities, higher alcohol and tobacco use, and malnutrition, as well as perhaps that this cohort is at least as similarly frail and comorbid as the octogenarians and older.²¹ Certainly, the cohort aged under 70 years who developed PJI in this study were very comorbid, and there is a strong selection bias as these patients were not deemed candidates for THA, which is the institutional norm for such patients. UTI at the time of index surgery was also an unexpected significant risk factor; although the PJI organism only correlated with the UTI in one case, it likely reflects a poorer physiological state, perhaps more susceptible to infection. Of particular note, taking antiplatelet or anticoagulants was not a risk factor for PJI in either analysis, nor was ASA grade. Generally, risk factors were nonmodifiable. While all identified significant predictors identified by this study were nonmodifiable, future research on the role of targeted screening for MDROs and UTIs and consideration of broader-spectrum empirical antibiotics may mitigate risk.^9,20

Unsurprisingly, the 90-day mortality was significantly worse in the HHA-PJI group when compared to both non-infected HHAs and eTHAs/TKAs (2.1%).²⁰ This difference was not found at 12 months or overall, suggesting that the PJI itself is the mortal threat, and in those who survive the initial septic insult and surgical management, they have similar longer-term survival as is seen in Figure 1. It is important to note that in the HHA-PJI group only eight patients survived beyond five years (1,825 days), which explains the divergence in the Kaplan Meier curve as time passes. Overall, the HHA mortality rates were similar to local Australian National Data on GHF morality in patients aged over 45 years (8.5% 30-day mortality; 15% at 90 days; and 26% at one year).²²

Overall, treatment success was low. Although this cohort is one of the largest in the literature, it is difficult to truly comment on the success of these 39 operatively managed patients who underwent 72 operations. Revision arthroplasty had the greatest success rate, and the lowest number, and likely the largest selection bias. Debridement with exchange of modular components appeared to outperform debridement without modular component exchange, a phenomenon seen in the eTHA/TKA setting. Debridement alone and primary excision arthroplasty were both more likely to fail than succeed. It is difficult to comment if the lower success seen with multiple surgeries reflects more difficult infections to manage, poor host factors, or technical factors to do with the surgeries themselves. However, in those who had revision arthroplasty after previous surgery, it was only successful in one-third of cases, suggesting earlier revision may be better for those patients for whom it is indicated.

Due to the unexpected multiorganism infection nature of the HHA-PJI at the time of the initial surgical management of the PJI, more than 50% of the patients were on an antibiotic regime that inadequately covered the bacteria eventually isolated. A key finding of this study, which potentially needs to be considered in future clinical practice, is related to the broader spectrum of pre-emptive antibiotic management when PJI develops after HHA following GHF.

This study’s limitations originate from its retrospective nature and the assumption that patients who underwent HHA would not move to another state. The influence of survival in this age group is substantial on our primary outcome and the impact of selection bias on treatment options is evident. Overall, the impact of this is the published PJI rate is likely lower than the real value. Adjusted multivariable analysis was limited by the relatively low number of PJIs in this cohort, to prevent the risk of overfitting and spurious inferences.²³ Large population-level epidemiological studies or well-designed data-linked registry studies are needed to reduce this underestimated rate. The strengths of this paper include the long follow-up period and its large population-based sample due to the well-defined two orthopaedic departments and one infection diseases service managing both the hip fractures and their surgical complications.

In conclusion, hip HHA-PJI is a different beast to eTHA/TKA-PJI. It occurs in at least 193/100,000 HHA/year and is associated with higher mortality. Risk factors include younger age, CKD, BMI > 35 kg/m², UTI at time of index surgery, and dementia, all non-modifiable. Unlike in the elective cohort, HHA-PJI presents more often with shock, draining sinuses, and dislocations. It is often polymicrobial and has a high rate of MDROs, bacteria often undertreated by conventional prophylaxis guidelines. The optimal surgical management is not known. Before meaningful research can be performed in this field, a consensus diagnostic criterion is a must. Following this, targeted research on identifying modifiable risk factors, investigating the optimum antibiotic prophylaxis in GHF and treatment algorithms based on more than intuition and borrowed lessons from the elective setting can be investigated.

Take home message

- Periprosthetic joint infection after hip hemiarthroplasty is a distinct, high-risk entity with poorer outcomes than elective arthroplasty infection.

- Further work is needed to identify strategies to prevent and manage these complications; in the meantime, realistic surgical goals and multidisciplinary management are critical to optimize results in this frail and comorbid cohort.

Data Availability

The datasets generated and analyzed in the current study are not publicly available due to data protection regulations. Access to data is limited to the researchers who have obtained permission for data processing. Further inquiries can be made to the corresponding author.