Keywords
total knee arthroplasty - human immunodeficiency virus - acquired immunodeficiency syndrome - postoperative complications - medical complications - surgical complications
With the average age of the United States population increasing, there is projected to be a large increase in arthritis cases to 67 million cases by 2030.[1] The most common cause of arthritis is osteoarthritis (OA), which currently affects 27 million adults.[2] The knee joint is the most affected joint; consequently,[3] total knee arthroplasty (TKA) is one of the most frequently performed surgical procedures. Patients diagnosed with human immunodeficiency virus (HIV) may be at additional risk for developing postoperative complications following TKA.[4]
[5] In the United States, over 1.1 million people are living with HIV, including an estimated 21% of whom are undiagnosed.[6] With the advent of highly active antiretroviral therapies (HAARTs), HIV-positive patients are living much longer than they had historically. As the average age for this patient population increases, the demand for TKA will increase due to age-associated joint degeneration.[7] This necessitates an understanding of how HIV affects perioperative and postoperative complications following TKA.
Multiple studies have attempted to analyze the postoperative complications of TKA in patients with HIV[7]
[8]
[9]; however, the literature is variable. Although some studies indicate that HIV-positive status alone is not a risk factor for increased complications following TKA,[7]
[10]
[11] other studies report that individuals with HIV may be at increased risk for perioperative complications.[4]
[5] Because some prior studies included subjects with both hemophilia and HIV, the data are subject to confounding.[12]
[13]
[14] Prior studies have also been limited by short-term follow-up and lack of subject stratification into HIV and acquired immunodeficiency syndrome (AIDS) cohorts,[15] which is an important distinction relative to the patient's immune status.
This study aimed to analyze the 90-day complications and 2-year postoperative surgical complications of patients with HIV who underwent TKA for OA. Our hypothesis was that patients with HIV, particularly patients with AIDS, would have significantly increased 90-day complication rates and higher risk of revision, particularly for prosthetic joint infection (PJI), at 2 years postoperatively.
Methods
Data Source and Collection
The PearlDiver Patient Records Database (www.pearldiverinc.com; Colorado Springs, CO) was used to perform a retrospective cohort analysis. Specifically, the Mariner dataset was utilized, including all-payer claim information for over 121 million patients from 2010 to 2019. Patients who underwent primary TKA were identified using International Classification of Diseases (ICD) procedure codes as well as Current Procedural Terminology (CPT) codes. Only patients with 2 years of follow up and who underwent primary TKA for knee OA were included in this study. The Mariner dataset provides active longitudinal follow-up based upon unique patient identifier codes that are not limited to changes in insurance status, minimizing loss to follow-up in the system. Therefore, patients with 2 years of follow up data included any patient with insurance continued coverage 2 years following TKA.
Exclusion and Inclusion Criteria
Patients with a diagnosis of HIV were identified using ICD 9/10 diagnosis codes. These patients were further subcategorized as having AIDS (ICD-9-D-042 and ICD-10-D-V08) or asymptomatic HIV (AHIV; ICD-9-D-B20 and ICD-10-D-Z21). The control group consisted of patients who underwent TKA for OA with no known diagnosis of HIV. Patients who were less than 18 years of age or had a staged or simultaneous bilateral TKA were excluded. Staged/bilateral TKAs were excluded to confirm that revision outcomes were from the primary TKA since CPT modifier codes are not present in the PearlDiver database. The flow chart documenting which PearlDiver patients were included in the present study is shown in [Fig. 1].
Fig. 1 Flow chart of PearlDiver patients included in the study.
Demographics/Comorbidities and Outcome Variables
Patient demographic and comorbidities observed included age, gender, and Elixhauser Comorbidity Index comorbidities[16] (congestive heart failure, arrhythmias, valvular disease, pulmonary circulatory disorders, peripheral vascular disease, hypertension, paralysis, other neuro disorders, chronic obstructive pulmonary disease, diabetes, hypothyroidism, chronic kidney disease, liver disease, peptic ulcer disease, lymphoma, metastatic cancer, nonmetastatic cancer, rheumatoid arthritis and cardiovascular disease, coagulopathy, fluid and electrolyte disorders, anemia due to blood loss, anemia due to deficiency, alcohol abuse, drug abuse, psychoses, depression, obesity, and smoking status). These were identified using ICD-9 and ICD-10 diagnosis codes. The primary outcomes of this study included 90-day postoperative complications: acute renal failure, surgical site infection (SSI), anemia, arrhythmia with and without atrial fibrillation, blood transfusion, bleeding complications, deep vein thrombosis (DVT), pulmonary thrombosis, stroke, heart failure, respiratory complications, pneumonia, urinary tract infection, sepsis, and death. The secondary outcome of this study was the 2-year complication rate, which included manipulation under anesthesia (MUA), all-cause revision, revision for peri-PJI, and revision for aseptic loosening.
Univariate/Multivariable Analysis
Patient data on demographics, comorbidities, and postoperative complications were analyzed using univariate and multivariable analysis via R software (https://www.R-project.org/, R Foundation for Statistical Computing, Vienna, Austria) provided by PearlDiver. Univariate analysis was conducted using Chi-squared and Student's t-tests where appropriate. Demographics and comorbidities with a p-value less than 0.2 on univariate analysis were included as independent variables in multivariable regression for postoperative complications. Results from the multivariable analysis were reported as odds ratios (ORs) and 95% confidence intervals (CIs). A p-value of 0.05 was used as the threshold level for significance.
Results
Study Cohort
In total, 855,373 patients who underwent TKA for a primary diagnosis of OA were included in this study. A total of 1,606 patients (0.19%) undergoing TKA for OA had a diagnosis of HIV at the time of their procedure. Of those, 1,338 patients (83.3%) had a diagnosis of AHIV at the time of TKA and 268 patients (16.68%) had a diagnosis of AIDS at the time of TKA. Demographics and comorbidities are detailed in [Table 1].
Table 1
Demographics and comorbidities of patients undergoing total knee arthroplasty
|
HIV
|
AHIV
|
AIDS
|
Control
|
|
Number
|
Percent (%)
|
p-Value
|
Number
|
Percent (%)
|
p-Value
|
Number
|
Percent (%)
|
p-Value
|
Number
|
Percent (%)
|
|
Total
|
1,606
|
|
|
1,338
|
|
|
268
|
|
|
853,767
|
|
|
Age (y)
|
58.76
|
|
<0.001
|
59.04
|
|
<0.001
|
57.31
|
|
<0.001
|
65.75
|
|
|
Gender
|
|
|
<0.001
|
|
|
<0.001
|
|
|
<0.001
|
|
|
|
Male
|
767
|
47.76%
|
|
627
|
46.86%
|
|
140
|
52.24%
|
|
312,607
|
36.62%
|
|
Female
|
839
|
52.24%
|
|
711
|
53.14%
|
|
128
|
47.76%
|
|
541,160
|
63.38%
|
|
CHF
|
359
|
22.35%
|
<0.001
|
310
|
23.17%
|
<0.001
|
49
|
18.28%
|
0.326
|
135,780
|
15.90%
|
|
Arrhythmias
|
730
|
45.45%
|
<0.001
|
606
|
45.29%
|
<0.001
|
124
|
46.27%
|
0.008
|
325,760
|
38.16%
|
|
Valvular disease
|
415
|
25.84%
|
<0.001
|
365
|
27.28%
|
<0.001
|
50
|
18.66%
|
0.159
|
191,524
|
22.43%
|
|
Pulmonary disorders
|
172
|
10.71%
|
<0.001
|
148
|
11.06%
|
<0.001
|
24
|
8.96%
|
0.562
|
66,743
|
7.82%
|
|
PVD
|
532
|
33.13%
|
<0.001
|
460
|
34.38%
|
<0.001
|
72
|
26.87%
|
0.109
|
192,772
|
22.58%
|
|
HTN
|
1,227
|
76.40%
|
<0.001
|
1,040
|
77.73%
|
<0.001
|
187
|
69.78%
|
0.501
|
613,115
|
71.81%
|
|
Paralysis
|
60
|
3.74%
|
<0.001
|
49
|
3.66%
|
<0.001
|
11
|
4.10%
|
0.041
|
18,108
|
2.12%
|
|
Other neuro disorders
|
206
|
12.83%
|
<0.001
|
170
|
12.71%
|
<0.001
|
36
|
13.43%
|
<0.001
|
62,413
|
7.31%
|
|
CPD
|
845
|
52.62%
|
<0.001
|
704
|
52.62%
|
<0.001
|
141
|
52.61%
|
<0.001
|
285,665
|
33.46%
|
|
Diabetes mellitus
|
791
|
49.25%
|
<0.001
|
690
|
51.57%
|
<0.001
|
101
|
37.69%
|
0.577
|
306,206
|
35.87%
|
|
Hypothyroidism
|
398
|
24.78%
|
0.019
|
345
|
25.78%
|
0.019
|
53
|
19.78%
|
0.002
|
245,347
|
28.74%
|
|
CKD
|
396
|
24.66%
|
<0.001
|
331
|
24.74%
|
<0.001
|
65
|
24.25%
|
0.012
|
154,788
|
18.13%
|
|
Liver disease
|
635
|
39.54%
|
<0.001
|
538
|
40.21%
|
<0.001
|
97
|
36.19%
|
<0.001
|
114,618
|
13.42%
|
|
PUD
|
58
|
3.61%
|
<0.001
|
49
|
3.66%
|
<0.001
|
9
|
3.36%
|
0.194
|
17,483
|
2.05%
|
|
Rheumatoid arthritis and CVD
|
365
|
22.73%
|
<0.001
|
317
|
23.69%
|
<0.001
|
48
|
17.91%
|
0.872
|
148,146
|
17.35%
|
|
Coagulopathy
|
243
|
15.13%
|
<0.001
|
198
|
14.80%
|
<0.001
|
45
|
16.79%
|
<0.001
|
63,431
|
7.43%
|
|
Fluid and electrolyte disorders
|
712
|
44.33%
|
<0.001
|
579
|
43.27%
|
<0.001
|
133
|
49.63%
|
<0.001
|
262,490
|
30.74%
|
|
Blood loss anemia
|
121
|
7.53%
|
<0.001
|
100
|
7.47%
|
<0.001
|
21
|
7.84%
|
0.065
|
43,999
|
5.15%
|
|
Deficiency anemia
|
442
|
27.52%
|
<0.001
|
382
|
28.55%
|
<0.001
|
60
|
22.39%
|
0.023
|
144,937
|
16.98%
|
|
Alcohol abuse
|
70
|
4.36%
|
<0.001
|
63
|
4.71%
|
<0.001
|
7
|
2.61%
|
0.011
|
8,000
|
0.94%
|
|
Drug abuse
|
484
|
30.14%
|
<0.001
|
396
|
29.60%
|
<0.001
|
88
|
32.84%
|
<0.001
|
50,616
|
5.93%
|
|
Psychoses
|
179
|
11.15%
|
<0.001
|
153
|
11.43%
|
<0.001
|
26
|
9.70%
|
<0.001
|
28,216
|
3.30%
|
|
Depression
|
886
|
55.17%
|
<0.001
|
717
|
53.59%
|
<0.001
|
169
|
63.06%
|
<0.001
|
285,586
|
33.45%
|
|
Obesity
|
678
|
42.22%
|
<0.001
|
563
|
42.08%
|
<0.001
|
115
|
42.91%
|
<0.001
|
252,169
|
29.54%
|
|
Smoking
|
629
|
39.17%
|
<0.001
|
509
|
38.04%
|
<0.001
|
120
|
44.78%
|
<0.001
|
108,159
|
12.67%
|
Abbreviations: BMI, body mass index; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disorder; CVD, collagen vascular disease; HTN, hypertension; PUD, peptic ulcer disease; PVD, peripheral vascular disease.
Note: All p-Value comparisons made between group of interest and control group; statistically significant p-values (p < 0.05) are indicated in bold.
Two-Year Surgical Complications
A total of 2.9% of patients with HIV (46 patients), of whom 85% (39 patients) had AHIV and 15% (7 patients) had AIDS, underwent revision TKA within 2 years of the initial procedure. This was significant compared with 1.8% of patients in the non-HIV cohort who required revision at this time point (p = 0.002). Patients with AHIV had a higher risk of PJI (p < 0.001), aseptic loosening (p = 0.011), and MUA (p < 0.001) than patients without HIV. Patients with AIDS were only at an increased risk for PJI (p = 0.002) compared with the control group. The breakdown of patients who had surgical complications is displayed in [Table 2].
Table 2
Two-year surgical complications following total knee arthroplasty
|
HIV
|
AHIV
|
AIDS
|
Control
|
|
Number
|
Percent (%)
|
p-Value
|
Number
|
Percent (%)
|
p-Value
|
Number
|
Percent (%)
|
p-Value
|
Number
|
Percent (%)
|
|
Total
|
1,606
|
|
|
1,338
|
|
|
268
|
|
|
853,767
|
|
|
Revision
|
46
|
2.86%
|
0.002
|
39
|
2.91%
|
0.004
|
7[a]
|
2.61%
|
0.460
|
15,552
|
1.82%
|
|
PJI
|
25
|
1.56%
|
<0.001
|
19
|
1.42%
|
<0.001
|
6[a]
|
2.24%
|
0.002
|
5,038
|
0.59%
|
|
Loosening
|
10[a]
|
0.62%
|
0.050
|
10[a]
|
0.75%
|
0.011
|
0
|
0.00%
|
0.705
|
2,704
|
0.32%
|
|
MUA
|
103
|
6.41%
|
<0.001
|
86
|
6.43%
|
<0.001
|
17
|
6.34%
|
0.084
|
34,748
|
4.07%
|
Abbreviations: MUA, manipulation under anesthesia; PJI, prosthetic joint infection.
Note: Groups were analyzed with chi-squared analysis compared with the control group. Statistically significant p-Values (p < 0.05) are indicated in bold.
a Per HIPAA requirements, <11 cohort sizes are not reportable.
However, on multivariable analysis, neither the AHIV nor the AIDS group was found to be at greater or lesser odds of 2-year revision surgery, PJI, aseptic loosening, or MUA compared with the control cohort. The ORs and CIs for all cohorts are shown in [Table 3].
Table 3
Multivariable analysis of 2-year complications and 90-day complications following total knee arthroplasty
|
HIV
|
AHIV
|
AIDS
|
|
Odds ratio
|
25%
|
75%
|
p-Value
|
Odds ratio
|
25%
|
75%
|
p-Value
|
Odds ratio
|
25%
|
75%
|
p-Value
|
|
2-year complications
|
|
Revision
|
0.842
|
0.618
|
1.119
|
0.257
|
0.872
|
0.622
|
1.186
|
0.405
|
–
|
–
|
–
|
–
|
|
PJI
|
1.237
|
0.807
|
1.805
|
0.298
|
1.139
|
0.695
|
1.748
|
0.578
|
1.639
|
0.643
|
3.397
|
0.236
|
|
Loosening
|
1.013
|
0.505
|
1.790
|
0.968
|
1.240
|
0.618
|
2.192
|
0.502
|
1.146
|
0.672
|
1.823
|
0.590
|
|
MUA
|
1.220
|
0.991
|
1.486
|
0.054
|
1.234
|
0.982
|
1.530
|
0.063
|
–
|
–
|
–
|
–
|
|
90-day complications
|
|
SSI
|
1.113
|
0.828
|
1.461
|
0.459
|
1.076
|
0.773
|
1.455
|
0.648
|
1.257
|
0.620
|
2.254
|
0.483
|
|
Renal failure
|
1.153
|
0.876
|
1.491
|
0.292
|
0.958
|
0.691
|
1.292
|
0.786
|
2.129
|
1.239
|
3.451
|
0.004
|
|
Anemia
|
–
|
–
|
–
|
–
|
0.697
|
0.587
|
0.821
|
<0.001
|
1.516
|
1.123
|
2.018
|
0.005
|
|
Arrhythmia w/ atrial fibrillation
|
0.814
|
0.668
|
0.984
|
0.037
|
0.813
|
0.655
|
1.000
|
0.055
|
–
|
–
|
–
|
–
|
|
Arrhythmia w/o atrial fibrillation
|
1.304
|
1.068
|
1.579
|
0.008
|
1.257
|
1.006
|
1.553
|
0.039
|
1.541
|
0.964
|
2.367
|
0.058
|
|
Bleeding complication
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
1.666
|
0.591
|
3.651
|
0.262
|
|
Blood transfusion
|
0.931
|
0.726
|
1.175
|
0.559
|
0.855
|
0.645
|
1.111
|
0.259
|
1.302
|
0.746
|
2.112
|
0.317
|
|
Pneumonia
|
1.318
|
0.946
|
1.783
|
0.087
|
1.094
|
0.736
|
1.559
|
0.638
|
2.473
|
1.299
|
4.265
|
0.003
|
|
Stroke
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
|
Death
|
1.179
|
0.466
|
2.415
|
0.689
|
1.151
|
0.410
|
2.500
|
0.755
|
–
|
–
|
–
|
–
|
|
DVT
|
1.085
|
0.813
|
1.416
|
0.563
|
1.054
|
0.765
|
1.413
|
0.736
|
1.210
|
0.596
|
2.176
|
0.561
|
|
Heart failure
|
0.912
|
0.689
|
1.188
|
0.505
|
0.885
|
0.652
|
1.178
|
0.416
|
–
|
–
|
–
|
–
|
|
Pulmonary embolism
|
1.232
|
0.832
|
1.757
|
0.273
|
1.173
|
0.756
|
1.737
|
0.450
|
–
|
–
|
–
|
–
|
|
Respiratory complication
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
|
Sepsis
|
1.294
|
0.843
|
1.891
|
0.209
|
1.301
|
0.813
|
1.962
|
0.239
|
1.206
|
0.370
|
2.863
|
0.713
|
|
UTI
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
–
|
Abbreviations: DVT, deep vein thrombosis; MUA, manipulation under anesthesia; PJI, prosthetic joint infection; SSI, surgical site infection; UTI, urinary tract infection; w/ atrial fibrillation, with atrial fibrillation; w/o atrial fibrillation, without atrial fibrillation.
Note: Groups were analyzed with chi-squared analysis compared with control group. Statistically significant p-Values (p < 0.05) are indicated in bold.
Ninety-Day Medical Complications
On univariate analysis, patients with AHIV were found to be at an increased risk of several minor 90-day complications such as SSIs (p < 0.001), arrhythmia without atrial fibrillation (p = 0.002), and pneumonia (p = 0.004) as well as major 90-day complications such as renal failure (p = 0.024) and sepsis (p < 0.001). Patients with AIDS were also at an increased risk of several 90-day complications including SSI (p = 0.007), anemia (p < 0.001), arrhythmia without atrial fibrillation (p < 0.001), bleeding complications (p = 0.024), DVT (p = 0.111), and pneumonia (p < 0.001). This cohort of patients was also at an increased risk of renal failure (p < 0.001) and sepsis (p = 0.187) within 90 days following TKA ([Table 4]).
Table 4
90-Day complications following total knee arthroplasty
|
HIV
|
AHIV
|
AIDS
|
Control
|
|
Number
|
Percent (%)
|
p-Value
|
Number
|
Percent (%)
|
p-Value
|
Number
|
Percent (%)
|
p-Value
|
Number
|
Percent (%)
|
|
Total
|
1,606
|
|
|
1,338
|
|
|
268
|
|
|
853,767
|
|
|
SSI
|
51
|
3.18%
|
<0.001
|
41
|
3.06%
|
<0.001
|
10
|
3.73%
|
0.007
|
13,056
|
1.53%
|
|
Renal failure
|
62
|
3.86%
|
<0.001
|
44
|
3.29%
|
0.024
|
18
|
6.72%
|
<0.001
|
19,808
|
2.32%
|
|
Anemia
|
222
|
13.82%
|
0.886
|
161
|
12.03%
|
0.089
|
61
|
22.76%
|
<0.001
|
116,699
|
13.67%
|
|
Arrhythmia with atrial fibrillation
|
130
|
8.09%
|
0.029
|
109
|
8.15%
|
0.055
|
21
|
7.84%
|
0.343
|
83,173
|
9.74%
|
|
Arrhythmia without atrial fibrillation
|
123
|
7.66%
|
<0.001
|
99
|
7.40%
|
0.002
|
24
|
8.96%
|
0.016
|
46,304
|
5.42%
|
|
Bleeding complication
|
14
|
0.87%
|
0.227
|
9[a]
|
0.67%
|
0.891
|
5[a]
|
1.87%
|
0.024
|
5,174
|
0.61%
|
|
Blood transfusion
|
73
|
4.55%
|
0.070
|
57
|
4.26%
|
0.277
|
16
|
5.97%
|
0.065
|
31,279
|
3.66%
|
|
Pneumonia
|
41
|
2.55%
|
<0.001
|
29
|
2.17%
|
0.004
|
12
|
4.48%
|
<0.001
|
10,749
|
1.26%
|
|
Stroke
|
18
|
1.12%
|
0.247
|
14
|
1.05%
|
0.464
|
4[a]
|
1.49%
|
0.387
|
7,064
|
0.83%
|
|
Death
|
6[a]
|
0.37%
|
0.088
|
5[a]
|
0.37%
|
0.133
|
1[a]
|
0.37%
|
0.941
|
1,434
|
0.17%
|
|
DVT
|
54
|
3.36%
|
0.001
|
44
|
3.29%
|
0.005
|
10[a]
|
3.73%
|
0.111
|
18,248
|
2.14%
|
|
Heart failure
|
64
|
3.99%
|
0.086
|
54
|
4.04%
|
0.097
|
10[a]
|
3.73%
|
0.748
|
27,316
|
3.20%
|
|
Pulmonary embolism
|
31
|
1.93%
|
0.014
|
25
|
1.87%
|
0.044
|
6[a]
|
2.24%
|
0.218
|
10,457
|
1.22%
|
|
Respiratory complication
|
10[a]
|
0.62%
|
0.930
|
9[a]
|
0.67%
|
1.000
|
1[a]
|
0.37%
|
0.822
|
5,736
|
0.67%
|
|
Sepsis
|
25
|
1.56%
|
<0.001
|
21
|
1.57%
|
<0.001
|
4[a]
|
1.49%
|
0.187
|
5,594
|
0.66%
|
|
UTI
|
100
|
6.23%
|
0.515
|
87
|
6.50%
|
0.310
|
13
|
4.85%
|
0.587
|
49,642
|
5.81%
|
|
Readmission
|
152
|
9.46%
|
<0.001
|
123
|
9.19%
|
<0.001
|
29
|
10.82%
|
<0.001
|
48,961
|
5.73%
|
|
LOS
|
2.80
|
|
<0.001
|
2.77
|
|
<0.001
|
4.60
|
|
<0.001
|
2.39
|
|
Abbreviations: DVT, deep vein thrombosis; LOS, length of stay; SSI, surgical site infection; UTI, urinary tract infection.
Note: Groups were analyzed with chi-squared analysis compared with control group. Statistically significant p-Values (p < 0.05) are indicated in bold.
a Per HIPAA requirements, <11 cohort sizes are not reportable.
On multivariable analysis, AHIV patients were found to have lower odds of anemia (OR: 0.697; 95% CI: 0.587–0.821; p < 0.001) compared with the control cohort while AIDS patients were found to have greater odds of anemia (OR: 1.516; 95% CI: 1.123–2.018; p = 0.005) relative to the control group. Additionally, AIDS patients were found to have greater odds of renal failure (OR: 2.129; 95% CI: 1.239–3.451; p = 0.004) within 90 days of TKA compared with the control group.
Discussion
It is imperative to understand the relationship between HIV/AIDS and postoperative complications following TKA given the increased life expectancy and rising incidence of TKA in this population. This study aimed to determine the 90-day complication rates and 2-year postoperative surgical complications in patients with a diagnosis of HIV who underwent TKA for OA. To our knowledge, no prior study has utilized a large administrative database to study the outcomes of HIV-positive patients who underwent OA-indicated TKA.
Patients with a diagnosis of HIV, both symptomatic and asymptomatic, had significantly increased odds of certain complications, including arrhythmia without atrial fibrillation and decreased odds of other complications, such as anemia, following TKA. However, patients with a diagnosis of AIDS had significantly increased odds of several 90-day postoperative complications including renal failure, anemia, and pneumonia (p = 0.004, p = 0.005, and p = 0.003, respectively). We were surprised to find that there were no significant differences in 2-year postoperative surgical complications between HIV-positive and HIV-negative TKA patients after multivariable regression analysis was performed ([Table 3]).
With the advent of HAARTs, HIV-positive patients are making significant improvements in survivorship,[17] which is expected to result in increased demand for total joint arthroplasties (TJAs) among HIV-positive patients.[18] There is currently no consensus in the literature regarding perioperative complications in HIV-positive patients undergoing TKA. Some studies report that HIV-positive patients are not at a significantly increased risk compared with the general population.[7]
[10]
[11] Other studies report that HIV-positive patients may be at a significantly increased risk for postoperative complications compared with their HIV-negative cohort.[4]
[5] O'Neill et al conducted a systematic review of 19 studies with a subgroup analysis of 4 studies.[9] The authors found that HIV-positive patients were at a significantly increased risk of infection and revision (risk ratio of 3.31 and 2.25, respectively). However, the authors caution that the quality of evidence was low and that there was heterogeneity between the studies. Parvizi et al report a significantly increased rate of early procedural failures among a cohort of HIV-positive patients undergoing TJA (13 knee procedures and 8 hip procedures).[4] Of the 21 total procedures, 12 were not successful due to loosening (average time of 9.5 years) and early repeat surgeries due to a decreased range of motion, hematoma of the knee, and other reasons. Six of the 12 procedural failures were due to infections, which had an average time to revision of 3.5 years. This study was limited due to its retrospective nature and small sample size. Olson et al conducted a multicenter retrospective cohort analysis of 110 HIV-positive patients and 240 control patients who underwent TJAs (85 total hip arthroplasties [THAs] and 25 TKAs).[19] The authors found a statistically significantly increased risk for venous thromboembolism in the HIV-positive TKA group compared with the control group (8.0 vs. 1.7% ; p = 0.046). Accordingly, they recommended that stronger prophylactic anticoagulation and a longer duration of prophylaxis be considered for HIV-positive patients undergoing TKA.
Boylan et al conducted a nationwide retrospective review and found that HIV-positive individuals undergoing TKA did not have significantly higher surgical complications overall compared with controls, but they were found to be at a significantly higher risk for developing infections.[20] Falakassa et al conducted a retrospective cohort analysis of 24 HIV-positive patients (31 THAs and one TKA).[7] In this study, there was one instance of aseptic loosening but no PJIs. The authors concluded that HIV-positive status did not place patients at an increased risk for postoperative PJIs and that patients with CD4 counts >200 on HAART therapy were at a similar risk to HIV-negative individuals.[7] Capogna et al reported no significant differences in infection, survivorship, or revision after TJA comparing 69 HIV-positive patients to 138 controls.[21]
Lin et al analyzed the Nationwide Inpatient Sample and identified 5,681,024 patients who were admitted for THAs and TKAs (0.14% were HIV-positive).[10] The authors found no significant increase in total complications between the HIV-positive and control group for TKA patients. While there were some increased complications (acute renal failure, wound infection, and increased indication for postoperative debridement and irrigation) identified in THA patients, it could not be concluded that HIV-positive status alone was a risk factor for increased complications after TJA. Roof et al conducted a multicenter retrospective study that analyzed discrepancies between 25 HIV-negative and 25 HIV-positive patients who underwent TKA.[8] The authors found that HIV-positive patients had an acceptable complication risk compared with HIV-negative patients (not statistically significant); however, HIV-positive patients who underwent TKA had significantly increased length of stay (3.8 vs. 2.28 days, p = 0.004).
While some 90-day complications were more prevalent in our HIV-positive cohort, there were no significant increases in 2-year postoperative surgical complications on multivariable linear regression analysis. Similar findings were seen in the study by Issa et al where no significant postoperative complications were seen with HIV-positive individuals who underwent TKA.[11] Knee implant survivorship between the HIV-positive and control groups was also comparable with no significant differences.
The significant heterogeneity in the literature indicates that HIV-positive status should not necessarily prevent patients from receiving TKAs. If proper preoperative and perioperative precautions are taken (use of antibiotics, anticoagulants, etc.), then HIV-positive patients may be successful TKA candidates. However, it is imperative that surgeons carefully weigh the risks and benefits of operating since some studies do report increased complications and infection risks for HIV-positive patients. Surgeons must assess each patient's individual risk factors, including comorbidities and past medical/surgical history, before deciding to move forward with TKA.
There are numerous strengths to this study. This study is unique in that it analyzes the 2-year postoperative surgical complications of patients with and without a diagnosis of HIV who underwent TKA for OA of the knee. The additional analysis comparing AHIV and AIDS is also novel. Study power was facilitated by using a large nationwide database, allowing the results to be more generalizable due to its inclusion of patient data from numerous institutions. However, the authors also recognize the limitations of this study. As with any retrospective database study, coding errors might have affected the results. Furthermore, we did not have access to the CD4 count or antiretroviral regimen each patient was taking, which may have influenced how the cohorts were identified. Further, revisions or other complications outside the 2-year study window were not captured. As such, it is likely that our study underestimates revision rates, and longer term studies are needed to better define the longitudinal risks of TKA in HIV-positive patients to guide patient counseling and shared decision-making.
Conclusion
After multivariable analysis, we found that patients with a diagnosis of HIV or AIDS are not at increased risk for developing surgical complications within 2 years of undergoing TKA. We recommend patient optimization focusing on the overall comorbidities rather than the patient's HIV status when considering TKA in HIV-positive patients.
