Outcomes of Total Joint Arthroplasty in HIV Patients

Abstract

Background

Advancement in human immunodeficiency virus (HIV) therapies has increased life expectancy. The need for joint replacement is expected to increase as this population develops degenerative changes from aging and avascular necrosis (AVN). Studies have shown a higher risk of peri-prosthetic joint infections (PJI) in HIV patients. However, these studies include a high percentage of hemophiliacs, which may be a confounding variable. With the advent of highly active anti-retroviral therapy (HAART) and evolving HIV demographics, we hypothesize the rate of PJIs in HIV patients are comparable to the general population.

Methods

We performed a retrospective cohort study using prospectively collected data from our arthroplasty database. We identified 24 HIV patients that underwent 31 primary hip and one primary knee arthroplasty between July 1, 2000 and September 30, 2012. Mean age was 50 years (range 31-74). Mean follow-up was 14 months (range 1.5-60).

Results

There were no PJIs in our HIV population. All HIV patients were non-hemophiliacs on HAART. Thirty-one total hip arthroplasties (THA) and one total knee arthroplasty were performed. Twenty-one HIV patients underwent THA for AVN. Eight patients had bilateral AVN. One patient needed revision for aseptic loosening. The mean CD4 count was 647 (194-1193). Mean viral load was undetectable in 19 patients and unavailable in five.

Conclusions

Our HIV population had a lower rate of PJI compared to infection rates in prior literature. Despite our limited patient population, our data suggests that well controlled HIV patients on HAART therapy with undetectable viral loads and CD4 >200 are at similar risk of PJI as the average population.

Introduction

According to the CDC, 1,148,200 people had a diagnosis of human immunodeficiency virus (HIV) in the United States at the end of 2010. The prevalence of HIV in the US continues to be high with an estimated 50,000 new infections in 2010¹. As the life expectancy in this population increases with improving therapies, patients will more frequently develop chronic conditions like degenerative joint disease in their hips and knees. In addition, this population has a 100-fold increased risk of developing avascular necrosis (AVN) of the femoral head². Both of these factors will contribute to the need for hip arthroplasty in this patient population.

Historically, HIV patients have been considered a high-risk population for musculoskeletal infections due to their compromised immune status. Infection in total joints can be detrimental, leading to multiple revision surgeries and poor outcomes. Current research concerning HIV infected individuals and total joint arthroplasty presents conflicting data. Some studies report complication rates as high as 57%⁵. The most common complications included deep infection and aseptic loosening. However, prior studies included predominantly HIV patients with hemophilia, which may be a confounding variable. The majority of these studies found that these infections occurred late and were related to repeated episodes of transient bacteremia from breaks in aseptic technique during factor transfusion. The demographics of HIV individuals have changed over the past several decades. With improved technologies in blood screening and recombinant factor, hemophiliacs represent a smaller proportion of HIV patients. The purpose of our study is to describe complications, particularly prosthetic joint infections (PJI), in non-hemophiliac HIV patients on HAART (highly active anti retroviral therapy).

Materials and Methods

We performed a retrospective chart review using prospectively collected data from our arthroplasty database. We identified 24 HIV-positive patients who underwent 31 primary hip and one primary knee arthroplasty between the dates of July 1, 2009 and September 30, 2012 at Jackson Memorial Hospital and the University of Miami Hospital. These charts were reviewed for: diagnosis, comorbidities, date of surgery, procedure performed, follow-up period, preoperative CD4 count, viral load, anti-retrovirals, history of smoking, alcohol, intravenous drug abuse, corticosteroid use, and postoperative complications (venous thromboembolism, infection, aseptic loosening).

Patients with underlying AVN were classified based on the Ficat grading system using an anterior-posterior pelvic radiograph that was evaluated by a board certified orthopaedic surgeon. All patients reviewed were graded as Ficat III or IV. A Ficat III grade was assigned to radiographs that demonstrated evidence of subchondral collapse, appearance of a sequestrum with a break on the articular margin and a normal joint space. A Ficat IV grade was assigned to radiographs that had the changes noted above in addition to joint space narrowing.

Infection prevention protocols were the same throughout the study period. Patients received cefazolin and vancomycin for prophylaxis unless they had a severe penicillin allergy. Allergic patients received clindamycin and vancomycin. Due to the high prevalence of methicillin resistant staph aureus (MRSA) infections at our institution, vancomycin administration is standard protocol for all arthroplasty patients. Antibiotics were administered within 1 hour of the procedure and for 24 hours postoperatively. All arthroplasties were performed in operating rooms equipped with vertical laminar flow with all members of the surgical team wearing helmet aspirator suits. Skin preparation included the use of alcohol and chlorhexidine lavage.

Postoperative wound management consisted of application of a sterile dressing, which was placed over the incision in the operating room and kept for 48 hours. The wound was inspected and the dressing changed using sterile gauze once daily. Patients were encouraged to contact their surgeon as needed or in the presence of fever, wound drainage, or any unexpected adverse events.

Results

In our series, we identified 24 patients with HIV that underwent 32 primary total joint arthroplasties (31 total hip arthroplasties (THA) and one total knee arthroplasty (TKA)) with a mean follow-up of 14 months (1.5-60 months) Table 1. Of our 24 patients, two were lost to follow-up. The remaining 22 patients all had minimum six week follow-up. The mean age of our population was 50 years old (range 31-74 years old). There were 17 males and seven females. All HIV infected individuals included in this study were on HAART therapy. Two patients had a history of intravenous drug use (IVDU). Eight patients had Hepatitis C and six patients had type II diabetes mellitus. The etiology was AVN of the hip in 29/31 (93.5%) THA. Of the 21 patients with AVN, five had a documented history of hyperlipidemia. Eleven hips with AVN were Ficat grade III, 16 hips were Ficat grade IV, and X-Rays for Ficat staging were not available for two hips. Of the 21 patients with AVN, eight patients had bilateral AVN. No patients had a history of corticosteroid use. Only one patient had a history of alcohol abuse. The remainders of the total joint arthroplasties (TJA) were performed for post-traumatic arthritis (one THA) and osteoarthritis (one TKA and one THA). There were no recorded infections at the surgical site. One HIV patient required a revision surgery for aseptic loosening. This was diagnosed one year after the initial operation and pre-operative infection work-up with ESR, CRP and hip aspiration were negative. Intraoperative cultures also revealed no evidence of infection. No postoperative dislocations, deep venous thrombosis, or superficial infection were noted. One patient experienced MSSA sepsis treated with oxacillin for two weeks from an infiltrated IV, which was diagnosed on postoperative day four. Peri-prosthetic infection was ruled as based on clinical exam and complete resolution of symptoms. His CD4 count was 165 at the time of bacteremia and 454 preoperatively. Another patient went into respiratory failure on postoperative day seven requiring tracheostomy and prolonged mechanical ventilation secondary to pneumonia and sepsis. His cultures results were Acinetobacter baumannii, Pseudomonas aeruginosa, Enterobacter, Klebsiella pneumoniae. His CD4 count dropped to 189 during this episode from his preoperative CD4 count of 425. This patient also had a documented history of Hepatitis C, alcohol abuse, alcoholic pancreatitis, 30 pack year history of smoking and squamous cell carcinoma of the larynx.

Table 1.

Patient Characteristics

	Sex/ Age	Diagnosis	Procedure	Ficat Stage	Co-Morbidities	IVDU	Follow-Up(Months)	Complications	CD4	Viral Load
1	M, 47	Hip AVN	THA	IV	DM	No	12	None	582	UD
2	M, 62	Hip AVN	THA	IV	HTN, Kaposi Sarcoma, Hyperlipidemia	No	10	None	765	UD
3	F.62	B Hip AVN	B THA	IV, IV	DM, HTN, CAD. Hyperlipidemia	No	18	None	1055, 1193	856/ 183
4	M, 47	B Hip AVN	B THA	III, IV	HCV C, HTN, Hyperlipidemia	NO	6	None	1034, 449	79/ 60
5	M,51	Hip AVN	THA	IV	HTN	No	2	None	659	UD
6	M,43	B Hip AVN	B THA	III,III	Toxoplasmosis	No	6	MSSA Bacteremia	373, 454	UD/UD
7	M,58	Hip AVN	THA	III	HCV, Laryngeal CA, TB, ETOH abuse	No	6	Pneumonia, Rapiratory Failure	425	UD
8	F, 58	Hip AVN	THA	III	HCV, HTN	NO	1.5	None	NA	NA
9	M, 61	Hip AVN	THA	IV	HCV, CVA	NO	LTF	None	290	UD
10	M,43	B Hip AVN	THA	IV, IV	None	No	5,2	None	557	UD
11	F, 54	Knee OA	TKA		Hep B, DM	No	13	None	1167	UD
12	M,50	Hip AVN	THA	IV	COPD	No	14	None	357	NA
13	M, 54	Hip Post-traumatic OA	THA		HCV, DM, cirrohsis, COPD, CKD, CAD	No	15	Aseptic loosening	664	UD
14	F, 56	OA	THA		HTN, DM, COPD	No	17	None	465	UD
15	M, 28	B Hip AVN	B THA	IV, IV	HCV	No	9, 12	None	NA, 390	NA, 23911
16	M, 31	Hip AVN	THA	III	None	No	LTF	None	NA	NA
17	F, 48	B Hip AVN	B THA	III, III	HCV, HBV, CAD	No	6	None	900	UD
18	M, 74	Hip AVN	THA	IV	Hyperpidemia	No	3	None	442	NA
19	M, 51	Hip AVN	THA	IV	None	No	6	None	>800, 797	UD/ UD
20	M, 36	B Hip AVN	B THA	NA	None	No	6O	None	NA	NA
21	F, 49	Hip AVN	THA	III	HCV, DM, COPD, Hyperlipidemia	Yes	6O	None	755	UD
22	M, 47	Hip AVN	THA	III	PCP pneumonia	NO	14	None	>700	UD
23	M, 31	B Hip AVN	B THA	IV, IV	PCP pneumonia	Yes	24	None	647	UD
24	F, 42	Hip AVN	THA	III	None	No	20	None	194	UD

*AVN (Avascular Necrosis), NA (not available), B (bilateral), OA (osteo-arthritis), THA (total hip arthroplasty), TKA (total knee arthroplasty), HCV (hepatitis C virus), HPV (Hepatitis B Virsus), HTN (hypertension), DM (diabetes mellitus), CAD (coronary artery disease), COPD (chronic obstructive pulmonary disease), PCP(pneumocystis carinii), LTF (lost to follow-up), CKD (chronic kidney disease), TB (tuberculosis), CA (cancer), ETOH (alcohol), UD (undectable), MSSA (methicillin sensitive staphylococcus aureus)

The mean CD4 count at time of surgery for our HIV patients was 646 (Range 194-1193). CD4 preoperative was not available for three patients. The mean viral load at time of surgery was undetectable in 19 patients. Two patients did not have viral load available. The remaining three viral loads were detectable, but <100 copies/mL.

Discussion

Assessing and predicting surgical outcomes for joint replacement surgery in HIV infected individuals has been a major challenge. Historically, studies have focused on joint replacement in HIV infected hemophiliac patients given this groups' high risk of transmission and early joint degeneration. However, hemophiliacs represent a technically challenging population, as they tend to be young, have associated angular deformities, contractures and poor bone quality. In addition, they have been shown to have higher complication rates, which make interpretation of these studies difficult.

Studies have shown conflicting data concerning PJI in the HIV positive hemophiliac population with rates ranging from 0-26.5%. A consensus regarding the safety of these elective procedures has been difficult to attain. Many studies have supported the use of preoperative CD4 counts to assess risks of PJI in HIV positive patients. In a retrospective review of 102 TJA in 73 HIV positive patients from eight hemophilia centers, Hicks et al. showed a correlation between a low CD 4 count and risk of infection⁴. The preoperative CD4 count was less than 0. 2 x10⁹/L³ in 62.5% (5/8) of the infected group, compared with 16.7% (7/42) in those not infected. Ragni et al. also showed an elevated rate of post-operative infections (15%) with CD4 counts below 0.2 x10⁹/L³. However, this study did not compare risk of PJI in HIV positive patients with CD4 counts greater than 0.2 x 10⁹/L³. Parvizi et al. found that average CD4 for infected TJA was 239+/112 compared to 523 +/- 171 in the non-infected group⁵.

Other studies have demonstrated CD4 counts to be unreliable for pre-surgical risk assessment. In a retrospective review of 55 total joint arthroplasties in 41 HIV positive patients Habermann et al. could not identify any differences when considering CD4 lymphocyte count⁶. Habermann also concluded that a difference in outcome between hemophilic patients with or without HIV could not be noted.

The majority (88.7%) of HIV disease transmission is now reported to occur through heterosexual and male- to-male sexual contact in otherwise healthy individuals⁷. Rather than joint degeneration from recurrent hemarthrosis, the HIV population of today is predominantly undergoing joint replacement for AVN of the hip². The exact pathophysiology of osteonecrosis in HIV patients remains unclear, however many predisposing factors have been described. Hyperlipidemia is a known complication of HIV infection and HAART, which may lead to induction of the atherosclerotic pathway leading to osteonecrosis. In our study, 5/21 patients with AVN also had a documented history of hyperlipidemia. Thus, further investigation on the use lipid lowering agents, such as statins, for prophylaxis against AVN in HIV patients may be warranted. There has also been a significantly higher incidence of AVN of the femoral head since the introduction of HAART. However, it is currently unclear whether this is due to HAART itself or the fact that HIV patients are now living longer to develop the late sequelae of symptomatic AVN.

To our knowledge, there are three recent papers published that evaluate TJA outcomes in the modern HIV population. Tornero et al. looked at 18 THAs in 13 HIV-infected patients and 36 THAs in 27 non-HIV-infected patients⁸. They found no significant differences in a variety of parameters including: mean time spent in surgery, the need for red cell transfusion, or the mean duration of hospitalization. The two groups showed similar postoperative functional results throughout the follow-up period (median 3.3 years in the HIV-positive group and 5.8 years in the HIV-negative group). In addition, they found no difference in the occurrence of complications.

Wang et al. reported on six HIV positive patients who underwent ten joint replacement procedures, including six total hip arthroplasties, two total knee arthroplasties, and one shoulder hemiarthroplasty. At an average follow-up period of 38.6 months, no infections were reported in their study⁹.

In the most recent published report by Capogna et al. on 57 HIV patients (69 TJA), a non-significant difference in early deep infection was found (4.4% v 0.72%). The HIV group, as a whole, had a higher incidence of methadone use and history of IVDU compared to the control group. Of the three deep infections in the HIV cohort, two had a history of IVDU. Cultures in two of the three infections in this group were MRSA. While not statistically different, there was 6.22 times increased odd of deep infection in the HIV group compared to the control group. This rate is much greater than the other two recent publications on this patient population. This may be related to the high prevalence of IVDU history in this study, as this is an established risk factor for deep infection in TJA.

Our HIV population had a lower rate of PJI compared to infection rates in prior literature. Similar to these studies, all our patients were on HAART therapy and the majority underwent TJA for AVN. No PJI or venous thromboembolisms were noted. However, our cohort did experience complications such as aseptic loosening requiring revision surgery, early line sepsis from an infiltrating IV, and postoperative pneumonia. Our patient who developed pneumonia both had a history of heavy smoking, laryngeal cancer and alcohol abuse. None of the other patients in our cohort had these identified risk factors. We feel that the risk factors of heavy smoking and alcohol abuse may play a more significant role in the predisposition to infection compared to a well-controlled history of HIV.

Our study is limited by small sample size, limited follow-up and retrospective design. Early studies have shown that the majority of infections in the hemophiliac population occur late due to repeated transfusions of factor needed in this population. This trend of late infection no longer appears to be the case as evidenced by the lack of late infections in Wang and Tornero's articles, which had mean follow-up periods greater than three years^8,9. While some infections may develop beyond our study period, a 12-month follow-up would be sufficient to identify a predisposition to early infection in this population. This is also supported by Pulido et al. study that identified that the majority of PJI (65%) was diagnosed within the first year after surgery¹². Poor follow-up is a difficult limitation to overcome in a county hospital setting with a patient population that is often insufficiently insured and has limited access to care. Capogna et al. experienced this point as well with only 75% patient follow-up at 180 days and 55% at one year¹⁰. Of our 24 patients, two were lost to follow-up. Of those who did follow-up, our average follow-up period was 14 months, which is where we would expect to see the majority of early PJI.

In the report by Capogna et al., of the postoperatively infected hip arthroplasties two-thirds of the patients grew MRSA in cultures. These patients only received Ancef for pre-operative antibiotics per their manuscript. Consideration should be given to adding vancomycin to this population for increased MRSA coverage. However, further studies are needed to determine the effect of adding vancomycin for preoperative prophylaxis¹⁰.

The benefit of TJA, even in the era of high HIV-related morbidity and mortality, was well recognized. Our study provides support to the growing body of literature that shows more acceptable infection rates in the modern HIV population undergoing TJA. As with any intervention, patient selection remains critical. Thus, further research on other markers such as CD4 and viral load for pre-operative risk assessment will be helpful for medical decision making.

Conclusions

Despite our limited patient population, our data suggests that well controlled HIV patients on HAART therapy with undetectable viral loads and CD4 >200 are at similar risk of PJI as the average population. This may be attributed to the fact that our HIV population is younger and closely medically managed. Caution should be exercised in the evaluation of elective joint 4. arthroplasty for HIV positive patients. Each patient must be considered individually and optimized in coordination with their HIV/AIDS specialist prior to surgery. Patient specific factors include: IVDU, CD4 count/ HIV viral load, and co-morbidities (diabetes, obesity, tobacco use, 5. alcohol use, Hepatitis C, inflammatory arthritides, etc.) Patients should be well informed of the increased risks and incidence of perioperative complications.

Contribution of Authors

Jonathan Falakassa MD: Litererature review, data gathering, data analysis, study conception and design, drafting article

Alejandro Diaz MD: Literature review, data gathering, 7. data analysis, and article revisions

Michaela Schneiderbauer MD: Principal Investigator, data analysis, and article revisions

Conflicts of Interest and Funding

In support of their research or preparation of this manuscript, the authors did not receive any grants or outside funding. None of the authors received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial 9. entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or non-profit organization with which the authors are affiliated or associated.