A Systematic Review and Meta-Analysis of Periprosthetic Joint Infection Rates in Morbidly Obese Patients Undergoing Total Hip Arthroplasty

Abstract

Introduction:

Periprosthetic joint infection (PJI) is a severe complication associated with higher rates in obese individuals after total hip arthroplasty (THA). Hard cutoffs for body mass index (BMI) levels may lead to restricted access to care; however, a certain level of obesity may warrant these restrictions for patient safety. The purpose of this study was to perform a systematic review and meta-analysis of articles comparing PJI rates in morbidly versus nonmorbidly obese patients undergoing primary THA.

Methods:

A systematic search of PubMed, EMBASE, Cochrane, and Google Scholar databases was conducted following Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines. Original studies comparing PJI rates in morbidly obese (BMI ≥40 kg/m²) and nonobese (BMI <40 kg/m²) THA patients were included. Data extraction, bias assessment, and quantitative synthesis were done.

Results:

After exclusion criteria, 10 studies comprising 46,080 THAs were included. Morbidly obese patients were found to have markedly higher rates of PJI compared with nonmorbidly obese patients (odds ratio = 4.332, 95% confidence interval [CI], 2.943 to 6.375, I² = 0, P-value = 0.901). Analysis of cohorts stratified by BMI showed consistent trends; morbidly obese patients demonstrated markedly increased risk of blood transfusions, postoperative fractures, superficial infections, dislocations, readmission within 90 days, surgical complications, and revision surgery.

Conclusion:

Morbidly obese patients demonstrate markedly increased risk of PJI and other complications after primary THA. This should be discussed with patients to safely provide the option of THA while minimizing restrictions on access to care.

Total hip arthroplasty (THA) is one of the most frequently performed orthopaedic procedures in the United States, projected to grow to 635,000 procedures annually by 2030.¹ The prevalence of obese patients in the arthroplasty population has grown alongside the national obesity epidemic, increasing from 7.0% to 22.7% from 2002 to 2017.² Furthermore, an increasing number of individuals have entered the morbidly obese category, or Class III obesity, as defined by the World Health Organization³ as a body mass index (BMI) of ≥40  kg/m². Previous studies have demonstrated that obesity may be a risk factor for poorer functional outcomes and higher rates of complications after THA.^4,5 These increased risks and costs have subsequently influenced some surgeons to use a BMI cutoff value of 40 kg/m², potentially restricting access to THA for a notable number of patients.⁶ Such cutoffs push the boundaries in regard to the ethical management of patients; however, this may be warranted and in the best long-term health interests of the patient.

Of the numerous possible complications associated with total joint arthroplasty, periprosthetic joint infection (PJI) is a devastating outcome with severe health and economic consequences.^7,8,9 Unfortunately, PJI is estimated to be responsible for 15% of all revision hip procedures and is associated with a 5-year mortality rate higher than that of several common malignancies.^10,11 It carries one of the heaviest burdens of arthroplasty complications and occurs at a markedly higher rate in obese patients.^5,12 However, the literature is conflicting on whether a BMI threshold exists for determining PJI risk within the obese population.^13,14 As such, the hard cutoff of BMI ≥40 kg/m² remains a point of debate and presents an opportunity for further investigation.

The purpose of this study was to perform a systematic review and meta-analysis of studies comparing PJI rates in morbidly obese patients versus nonmorbidly obese patients undergoing primary THA. We hypothesized that morbidly obese patients would demonstrate higher rates of PJI, complications, and adverse events compared with patients with BMI <40 kg/m².

Methods

Search Strategy and Article Selection

The study identification and selection process was conducted in accordance with the 2009 guidelines for Preferred Reporting Items for Systematic Reviews and Meta-analyses presented¹⁵ in Figure 1. The following databases were searched for original articles published before May 2023: PubMed, EMBASE, Cochrane Central Register of Controlled Trials, and Google Scholar. The following search terms were used: (BMI OR body mass index OR obesity OR obese) AND (periprosthetic joint infection) AND (THA OR total hip arthroplasty). All articles were evaluated with no additional restrictions.

Figure 1.

Flow chart showing Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guideline.

Two independent reviewers (initials hidden for review purposes) evaluated all abstracts of identified articles for agreement with the following inclusion criteria: (1) studies presenting original data; (2) studies available in English; (3) studies comparing PJI outcomes in morbidly obese patients with nonmorbidly obese patients undergoing primary THA; (4) level of evidence (LOE) of I to IV as described by Sackett et al.¹⁶ If reports meeting inclusion criteria had overlapping study groups, the article reporting the largest patient population was included in final extraction, and the remaining were excluded.

The following exclusion criteria were applied during article selection: (1) preclinical or biomechanic articles; (2) technical notes; (3) editorial articles; (4) review articles; (5) articles excluding patients with osteoarthritis; and (6) articles reporting on procedures unrelated to THA. Full-length texts were obtained when abstracts were insufficient for screening purposes. References of all included articles were reviewed to ensure that all relevant studies were captured.

Four hundred six articles were identified after searching PubMed (n = 231), Embase (n = 160), and Cochrane (n = 15). After 84 duplicates were removed, 322 studies were screened. Of these, 146 were excluded due to inclusion of revision surgeries (n = 55), study designs that were systematic reviews, narrative reviews, or abstracts (n = 47), inclusion of knee or shoulder surgeries (n = 18), limited article access (n = 12), inclusion of fracture or resection arthroplasty (n = 11), and non-English language (n = 3). Next, 166 studies were excluded due to lack of a study cohort with BMI ≥ 40 kg/m² (n = 144), lack of BMI stratification (n = 14), combination of THA and total knee arthroplasty data (n = 4), preoperative diagnosis that did not include osteoarthritis (n = 2), or the study being of insufficient LOE (ie, LOE V) (n = 2). This left 10 studies to be included in this systematic review and meta-analysis.

Data Extraction, Synthesis, and Outcomes Measurement

Two independent reviewers (initials hidden for review purposes) extracted the following variables from each included study, as available: characteristics of the study population (including sample size, age, sex, and length of follow-up), patient diagnoses, interventions done, perioperative data, and postoperative complications, clinical and functional outcomes scores, and publication characteristics (including journal and year of publication, study design, and LOE).

Assessment of Risk of Bias and Quality of Evidence

The methodological quality of the studies included was evaluated by two of the authors (initials hidden for review purposes) presented in Figure 2. The Methodological Index for Non-Randomized Studies criteria was used to assess the methodological quality of nonrandomized studies, such as case-control and cohort studies.¹⁷ Noncomparative studies are assessed with a maximum score of 16, whereas comparative studies have a maximum score of 24. Higher Methodological Index for Non-Randomized Studies scores are representative of greater methodological quality. Discordance between the reviewers was settled by a third, independent reviewer (initials hidden for review purposes). Interrater reliability is represented by Cohen κ coefficient using SPSS version 27.0 (IBM).^18,19

Figure 2.

Chart showing Methodological Index for Non-Randomized Studies (MINORS) criteria quality assessment analysis.

Quantitative Synthesis of Outcomes

Noncomparative data were reported using descriptive statistics. For comparative data, the pooled event rate with 95% confidence interval (CI) of each population of patients was compared for PJI incidence. For each population of patients, the standardized treatment effect was calculated using standardized mean difference, with a 95% CI for postoperative patient-reported outcome scores. Complications that had quantitative data available from three or more articles, which included rates of superficial infection, blood transfusion, periprosthetic fracture, and hip dislocation, were chosen for the meta-analysis. SMDs and pooled event rates were analyzed using random-effects modeling to generate forest plots and account for the anticipation of heterogeneous populations included in the analysis.²⁰ Heterogeneity was evaluated using I² statistics.²¹ A P ≤ 0.05 was considered statistically significant. All analyses of SMD and event rate computations with a 95% CI were done using SPSS version 27.0 (IBM).¹⁹

Results

Study Characteristics

Of the 10 included studies, nine were retrospective cohort studies (LOE III) and one was a prospective cohort study (LOE II). A total of 46,080 primary THAs were included, in which 6,094 were patients with a BMI ≥ 40 kg/m². Patients with a BMI ≥ 40 kg/m² were markedly younger than those with a BMI < 40 kg/m², with the cohorts having mean ages of 62.87 and 65.84 years, respectively. Morbidly obese patients also underwent markedly longer surgeries but had markedly shorter lengths of stay (Table 1). The characteristics of the morbidly obese patients from each study are described^22,24-29,31 in Table 2. Other complications reported in the included articles are shown in Table 3.

Table 1.

Characteristics of Included Patients

Factor or Variable	BMI < 40	BMI > 40
Total patients	39,986	6094
Mean age (yrs)	65.84 ± 19.78	62.87 ± 14.24a
Female/male (%)	59.1/40.9	62.5/37.5
Mean BMI	25.82 ± 9.77	43.80 ± 4.83
Mean length of surgery (minutes)	90.83 ± 16.57	111.69 ± 13.13a
Mean length of stay (d)	4.10 ± 1.99	3.18 ± 1.05a

BMI = body mass index

^aNotable difference (P < 0.05) between the two cohorts.

Table 2.

Characteristics of Studies Evaluating Morbid Obesity and Periprosthetic Joint Infection Development After Total Hip Arthroplasty

First Author (Year)	Study Design (LOE)	Patients (BMI > 40)	Treatment Approach	BMI > 40 Mean Age	BMI > 40 Sex F/M (%)	Mean Length of Surgery (mins) (BMI > 40)	Mean Length of Stay (d) (BMI > 40)	PJI Rate (%) (BMI > 40)
Shevenell et al22 (2023)	Retrospective cohort study (III)	341	ABMS	61.3	65.4/34.6	73.5	1.5	0.29
Nizam et al23 (2022)	Retrospective cohort study (III)	6	DAA (noncemented)	68	64.7/35/3	NR	NR	0
Ryan et al24 (2022)	Retrospective cohort study (III)	88	NR	60	68/32	122	NR	0.02
Argyrou et al25 (2022)	Retrospective cohort study (III)	86	DAA (noncemented)	64.9	54.7/45.3	70.12	NR	1.2
Luger et al26 (2021)	Retrospective cohort study (III)	19	Anterolateral (noncemented)	65.17	57.9/42.1	85.36	15.21	10.5
Abdulla et al27 (2020)	Retrospective cohort study (III)	811	NR	61.3	58.1/41.9	NR	4.3	OR = 8.51
Matar et al28 (2020)	Retrospective cohort study (III)	3635	NR	63	63.3/36.7	116	3	2.8
Hartford et al29 (2020)	Retrospective cohort study (III)	96	DAA	62.3	57.3/42.7	115.9	2.81	2.1
Wagner et al30 (2016)	Retrospective cohort study (III)	994	NR	NR	NR	NR	NR	HR = 4.6
Andrew et al31 (2008)	Prospective cohort study (II)	18	Anterolateral or posterior (noncemented)	60.6	72.2/27.8	117.2	8.9	0

ABMS = anterior-based muscle sparing; BMI = body mass index; DAA = direct anterior approach; F/M = female/male; HR = hazard ratio; LOE = level of evidence; NR = not recorded; OA = osteoarthritis; OR = odds ratio; PJI = periprosthetic joint infection; THA = total hip arthroplasty

Table 3.

Other Complications After Total Hip Arthroplasty for Morbidity Obese Patients

First Author (Year)	Blood Transfusion (%) (BMI > 40)	Fracture (%) (BMI > 40)	Superficial Infection (%) (BMI > 40)	Dislocation (%) (BMI > 40)	Readmission within 90 days (%) (BMI > 40)	Surgical Complication (%) (BMI > 40)	Revision (%) (BMI > 40)
Shevenell et al22 (2023)	0.88	0.88	1.17	0	3.81	NR	NR
Nizam et al23 (2022)	NR	0	0	0	NR	NR	NR
Ryan et al24 (2022)	NR	NR	0.03	0	NR	NR	NR
Argyrou et al25 (2022)	NR	NR	8.1	NR	NR	16.3	NR
Luger et al26 (2021)	10.5	0	NR	0	NR	15.8	10.5
Abdulla et al27 (2020)	OR = 0.38	NR	NR	NR	OR = 1.59	NR	NR
Matar et al28 (2020)	NR	NR	NR	0.6	0.5	3.6	2.4
Hartford et al29 (2020)	7.8	5.2	NR	0	4.2	16.6	NR
Wagner et al30 (2016)	NR	NR	HR = 4	HR = 1.4	NR	NR	HR = 1.5
Andrew et al31 (2008)	NR	NR	NR	5.6	NR	NR	0

ABMS = anterior-based muscle sparing; BMI = body mass index; DAA = direct anterior approach; F/M = female/male; HR = hazard ratio; LOE = level of evidence; NR = not recorded; OA = osteoarthritis; OR = odds ratio; PJI = periprosthetic joint infection; THA = total hip arthroplasty

Of the included studies, one study reported 30-day outcomes,²⁷ two reported 90-day outcomes,^26,29 and two reported 1-year minimum follow-up.^23,28 Andrew et al³¹ reported 5-year minimum follow-up. Ryan et al and Argyrou et al both reported 2-year minimum follow-up,^24,25 while Wagner et al³⁰ reported a 7.5-year mean follow-up. One study did not report follow-up data.²²

Overall Meta-Analysis: Periprosthetic Joint Infection Rates after Total Hip Arthroplasty in Body Mass Index ≥ 40 kg/m² vs. Body Mass Index <40 kg/m²

Nine articles had sufficient quantitate data to perform meta-analysis comparing morbidly obese patients with those with BMI <40 kg/m².^22-29,31 Meta-analysis revealed that the risk of PJI after primary THA was markedly increased in morbidly obese patients (odds ratio [OR] 3.7, 95% CI, 2.7 to 4.9), with low heterogeneity among the studies (I² = 0%, P = 0.505; Figure 3).

Figure 3.

Forest plot showing meta-analysis results of the correlation between body mass index (BMI) > 40 kg/m² vs. <40 kg/m² and periprosthetic joint infection (PJI) after total hip arthroplasty (THA).

Revision surgery and Revision Rates after THA in BMI ≥ 40 kg/m² vs. Body Mass Index < 40 kg/m²

Four articles had sufficient data for meta-analysis of revision surgery^24,25,26,29 and revision^24,25,28,31 rates between patients with BMI ≥ 40 kg/m² vs. those with BMI <40 kg/m². Morbidly obese patients had markedly increased risk for both revision surgery (OR 3.5, 95% CI, 1.6 to 7.9) and revision following primary THA (OR 2.7, 95% CI, 1.8 to 3.9) compared with patients with BMI < 40 kg/m² (Figures 4 and 5).

Figure 4.

Forest plot showing meta-analysis results of the correlation between body mass index (BMI) > 40 kg/m² vs. <40 kg/m² and reoperation surgery after total hip arthroplasty (THA).

Figure 5.

Forest plot showing meta-analysis results of the correlation between body mass index (BMI) > 40 kg/m² vs. <40 kg/m² and revision after total hip arthroplasty (THA).

Other Complication Rates after THA in BMI ≥ 40 kg/m² vs. BMI<40 kg/m²

Morbidly obese patients had markedly increased risk for superficial infection (OR 12.5, 95% CI, 4.3 to 36.6) following primary THA compared with patients with BMI < 40 kg/m² (Figure 6). No significant difference was found between the two groups for rates of blood transfusion, periprosthetic fracture, or hip dislocation (Figures 7–9).

Figure 6.

Forest plot showing meta-analysis results of the correlation between body mass index (BMI) > 40 kg/m² vs. <40 kg/m² and superficial infection after total hip arthroplasty (THA).

Figure 7.

Forest plot showing meta-analysis results of the correlation between body mass index (BMI) > 40 kg/m² vs. <40 kg/m² and need for blood transfusion during total hip arthroplasty (THA).

Figure 9.

Forest plot showing meta-analysis results of the correlation between body mass index (BMI) > 40 kg/m² vs. <40 kg/m² and hip dislocation after total hip arthroplasty (THA).

Figure 8.

Forest plot showing meta-analysis results of the correlation between body mass index (BMI) > 40 kg/m² vs. <40 kg/m² and periprosthetic fracture after total hip arthroplasty (THA).

Discussion

The main findings of this study are that morbidly obese patients face a markedly greater risk of developing a PJI after undergoing primary THA compared with nonmorbidly obese patients. To our knowledge, this is the first systematic review of its kind to investigate morbid obesity (BMI > 40 kg/m²) as a primary risk factor for PJI after THA. Furthermore, it includes the largest sample size of morbidly obese patients undergoing THA to date.

Morbidly obese patients demonstrated an overall 3.65 times greater risk of developing PJI after THA in comparison to nonmorbidly obese patients. Furthermore, morbidly obese patients face greater risks in the acute postoperative period, having higher rates of superficial infection, and requiring revision surgery more frequently compared with their nonmorbidly obese counterparts. Previous literature has established obesity as a prominent risk factor for the development of infection following total joint arthroplasty. A meta-analysis done by Guo et al³² assessed reinfection rates after two-stage revision surgery for obese patients who underwent either a THA or total knee arthroplasty and found that obese patients had markedly higher rates of reinfection after two-stage revision. In addition, Kong et al³³ found that among multiple risk factors, obese patients who underwent total joint arthroplasty (TJA) were 1.54 times more likely to develop PJI. However, morbidly obese patients were not included, and PJI was not the primary outcome of the study.³³ This study supplements the current body of literature, suggesting that morbid obesity presents an even greater risk factor for infection after THA than nonmorbid obesity.

Along with PJI, previous studies have investigated the relationship between obesity and other perioperative and postoperative complications in patients undergoing TJA. In a study conducted by Onggo et al,³⁴ the authors discovered that obese patients are at a higher risk for developing infection or requiring revision after THA. Although Onggo et al compared obese versus nonobese patients, their findings reinforce the similar results of this study. Ward et al³⁵ similarly used the Veterans Affairs Surgical Quality Improvement Program database to analyze surgical complications after TJA for patients with morbid obesity and found a similar increase in revision surgery and superficial infection rates and increased rates of acute kidney injury, cardiac arrest, and 1-year mortality. Complications are largely related to a more challenging surgical exposure in morbidly obese patients, which is associated with increased surgical time, greater blood loss, and risk of implant malposition.^36,37,38 Ashkenazi et al³⁹ recently demonstrated that these complications may be mitigated when treated by high-volume surgeons who encounter more morbidly obese patients. Future studies investigating the learning curve of THA in morbidly obese patients would be warranted to further explore this question.

Interestingly, we did find that morbidly obese patients were markedly younger than their nonmorbidly obese counterparts when having TJAs performed, which may be a positive indicator for their prognosis. However, this finding may also just reflect the earlier development of end-stage osteoarthritis in this cohort or patient selection practices in which surgeons are hesitant to operate on patients who are older and morbidly obese.

Our findings emphasize the need for physicians to communicate clearly about the increased risk of PJI and other complications associated with morbid obesity and to try to optimize the health status of their patients, but it also has implications for informing future policy regarding arthroplasty care in complex, higher-risk patients. With growing evidence demonstrating that comparative reimbursement is lower among sicker patients compared with their healthier counterparts secondary to longer surgical times, increased lengths of stay, and poorer discharge disposition, an increasing proportion of surgeons shy away from patients they deem to be higher risk, inevitably intensifying the disparities within arthroplasty access.^40,41 Exploring strategies that include rather than shut out morbidly obese or sicker patients will be essential in incentivizing surgeons and healthcare institutions in providing care in a country where rates of obesity continue to climb. Future research needs to be done to identify comorbidities that may be optimized within the morbidly obese population to improve the safety profile of arthroplasty in these patients. Questions remain if BMI cutoffs should be adjusted to be higher or lower than 40 to perform arthroplasty safely in the setting of other comorbidities.

This study has several limitations. Although there was low to no heterogeneity among our cohorts, preoperative diagnoses and indications for THA varied among the studies. Follow-up time also varied among studies, and thus, our systematic review may not capture complications that occur in the mid- or long-term period. Furthermore, there was a mix of different surgical techniques and approaches used for the THA procedures. Such differences may have influenced complications and PJI outcomes after receiving a THA. In addition, the variability in defining other BMI cohorts limits the study's ability to compare morbidly obese outcomes with other BMI groups (i.e., overweight).

Conclusion

Morbidly obese patients are 3.7 times more likely to develop PJI after THA and are 3.5 times more likely to require revision surgery compared with patients with lower BMI. These substantially higher risks should be considered and discussed before undergoing the procedure. In improving access to arthroplasty care, future studies should explore modifications on both the surgeon and patient side that can mitigate poor outcomes in morbidly obese patients.