Operative Time and Risk of Surgical Site Infection and Periprosthetic Joint Infection: A Systematic Review and Meta-Analysis

Abstract

Background

The purpose of this study was to perform a systematic review and meta-analysis on the association between operative time and peri-prosthetic joint infection (PJI) after primary total hip arthroplasty (THA) and total knee arthroplasty (TKA).

Methods

PubMed, Embase, and Cochrane CENTRAL databases were searched for relevant articles dating 2000-2020. Relationship of operative time and PJI rate in primary total joint arthroplasty (TJA) was evaluated by pooled odds ratios (OR) and 95% confidence intervals.

Results

Six studies were identified for meta-analysis. TJA lasting greater than 120 minutes had greater odds of PJI (OR, 1.63 [1.00-2.66], p=0.048). Similarly, there were greater odds of PJI for TJA procedures lasting greater than 90 minutes (OR, 1.65 [1.27-2.14]; p<0.001). Separate analyses of TKA (OR, 2.01 [0.76-5.30]) and THA (OR, 1.06 [0.80-1.39]) demonstrated no difference in rates of PJI in cases of operative time ≥ 120 minutes versus cases < 120 minutes (p>0.05 for all). Using any surgical site infection (SSI) as an endpoint, both TJA (OR, 1.47 [1.181.83], p<0.001) and TKA (OR, 1.50 [1.08-2.08]; p=0.016) procedures lasting more versus less than 120 minutes demonstrated significantly higher odds of SSI.

Conclusion

Following TJA, rates of SSI and PJI are significantly greater in procedures ≥120 minutes in duration relative to those < 120 minutes. When analyzing TKA separately, higher rates of SSI were observed in procedures ≥ 120 minutes in duration relative to those <120 minutes. Rates of PJI in TKA or THA procedures alone were not significantly impacted by operative time.

Level of Evidence: V

Introduction

Periprosthetic joint infection (PJI) is a severe complication of total joint arthroplasty (TJA) procedures for both the healthcare team and patient.¹ As the annual volume of total knee and total hip arthroplasty procedures increases, so does the prevalence of revision arthroplasty secondary to surgical site infections (SSI) and PJI.² Operative time has previously been evaluated as a potential independent risk factor for development of postoperative SSI and/or PJI.^3-8 While many studies have suggested that longer operative times may increase the risk of postoperative SSI and PJI,^4-16 other studies have not demonstrated increased risk of SSI and PJI in cases with longer operative times.^3,17,18 In light of these competing results, establishment of specific duration thresholds and/or time increments that may place patients at increased risk for infectious complications is needed. The purpose of this study was to perform a systematic review and meta-analysis to assess potential associations qualitatively and quantitatively between specific thresholds of operative time, SSI, and PJI following primary TJA.

Methods

Search strategies were developed with the assistance of a health sciences librarian with expertise in conducting systematic searches. The strategies were developed by the authors and a health sciences librarian by gathering, evaluating, and testing search terms. Comprehensive search strategies, including searching by index and keyword methods, were devised for the following databases: PubMed, Embase (Elsevier platform), and Cochrane CENTRAL (Wiley). To maximize sensitivity, no pre-established database filters were used other than an English language filter. Searches were finalized in October 2019 and then updated in October 2020 to identify results published during the systematic review process. The full PubMed search strategy, as detailed below in Table 1, was adapted for use with the other electronic databases. Complete search strategies are available upon request. In addition to database searches, references and citing papers of relevant papers were reviewed.

Table 1.

Final PubMed Search October 2019

#1 “Operative Time”[Mesh] OR “Time Factors”[Mesh] OR “operative time”[Title/Abstract] OR “operating time”[Title/Abstract] OR “operative duration”[TItle/Abstract] OR “operation duration”[Title/Abstract] OR “surgical time”[Title/Abstract] OR “surgery time”[Title/Abstract] OR “surgery duration”[Title/Abstract] OR “surgical duration”[Title/Abstract]

#2 “Arthroplasty, Replacement, Hip”[Mesh] OR “Arthroplasty, Replacement, Knee”[Mesh] OR total hip[Title/Abstract] OR THA[Title/Abstract] OR thr[Title/Abstract] OR total knee[Title/Abstract] OR tka[Title/Abstract] OR tkr[Title/Abstract] OR tja[Title/Abstract] OR “hip prosthesis implantation”[Title/Abstract] OR total joint arthroplasty[Title/Abstract] OR partial hip arthroplasty[Title/Abstract] OR partial knee arthroplasty[Title/Abstract] OR “hip hemiarthroplasty”[Title/Abstract] OR unicom-partmental knee arthroplasty[Title/Abstract]

#3 “Postoperative Complications”[Mesh:noexp] OR “Surgical Wound Dehiscence”[Mesh] OR “Infections”[Mesh] OR ssi[Title/Abstract] OR surgical site infection[Title/Abstract] OR joint infection[Title/Abstract] OR pji[Title/Abstract] OR postoperative complications[Title/Abstract] OR postoperative complication[Title/Abstract] OR post operative complication[Title/Abstract] OR post operative complications[Title/Abstract] OR post operative infection[Title/Abstract] OR post operative infections[Title/Abstract] OR postoperative infection[Title/Abstract] OR postoperative infections[Title/Abstract] OR wound complication[Title/Abstract] OR wound complications[Title/Abstract]

#4 “Editorial”[Publication Type] OR “Comment”[Publication Type] OR “Letter”[Publication Type] OR “Review”[Publication Type]

(#1 AND #2 AND #3) NOT #4 Limited to English[lang]=1433

After removal of duplicates, a total of 2,577 records were obtained. Titles and abstracts of all obtained records were screened by three independent reviewers (N.M.S., C.N.C., N.A.B.) to assess if the relationship between operative time and infection were evaluated. The three reviewers compared screening results and reached an agreement on which records to include for full-text analysis. Inclusion of articles into qualitative synthesis was based upon the following criteria: (1) relationship between infection and operative time was evaluated in primary THA and/or TKA procedures, (2) any wound complication, PJI, or SSI was defined as an endpoint, (3) operative time was reported in a manner sufficient for analyzing effect on infection rate. An additional inclusion criterion for quantitative synthesis was: (4) provision of sufficient data for calculating pooled odds ratios (OR) with a 95% confidence interval.² Non-primary THA/TKA procedures, editorials, commentaries, letters, reviews, and non-English language studies were excluded from analysis. To avoid overlapping data, only one study utilizing the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database was included;⁷ all other studies utilizing ACS-NSQIP were excluded. The selected study by Duchman et al.⁷ was included as it classified operative time in a manner that best matched our specific-increment analysis. Exclusion of full-text articles were recorded and are shown in Figure 1.

Figure 1.

PRISMA 2009 Flow Diagram

One-hundred sixty-eight full-text articles were reviewed for inclusion. These studies were then evaluated independently by three reviewers, and important characteristics were recorded for quality assessment. The characteristics included: author, title, year published, years of study, institution, country, study design, methods of controlling for cofounders, controls (gender, body mass index (BMI), American Society of Anesthesiologists (ASA) classification, length of stay, operative time, etc.), number of THA, TKA, and TJA procedures in each respective cohort, minimum follow-up, mean follow-up, definition of operative time, mean operative time, distribution of operative time intervals (if any) and definition/criteria of wound complications, SSI and/or PJI.

After reviewer consultation, data from seventeen studies were extracted for qualitative synthesis and six studies for quantitative synthesis, detailed in Table 2. Definitions of operative time varied across studies; for the purpose of the present study, existing studies were placed into one of two categories of operative time: (1) skin incision to wound closure, and (2) total anesthesia time. All studies included in the quantitative analysis defined operative time as skin incision to wound closure. There was significant heterogeneity in definitions of wound complications between studies. To compensate for this, we defined an SSI as any reported description of surgical site infection, including PJI. If studies specifically reported PJI, an additional analysis with only PJI as an endpoint was performed. Follow-up time ranged from two weeks to four years post operation across studies; given this variability, we were unable to control for an average follow-up time in our analysis, and no study exclusions were made based on the study follow-up period. Distribution of operative time categories varied significantly between studies. For the present study, we established the following paired categories of operative time for use in qualitative and quantitative analysis: (1): <90 minutes or ≥ 90 minutes and (2) <120 minutes or ≥ 120 minutes.

Table 2.

Meta-Analysis Study Characteristics

Author	Country	Minimum, Average Follow Up	# THA	# TKA	# TJA	Operative Time Definition	Operative Time Categorization	Definition SSI	Definition PJI
Aggarwal	United States	90 days, 90 days	6,086	0	6,086	n/a	>120 minutes	n/a	NHSN deep SSI definition
Anis	United States	Minimum n/a, 2 ±1 year	0	11,840	11,840	Skin incision to wound closure	<85, 85-102, 103-121, >121 minutes	Skin or superficial wound infections treated with antibiotics alone	Deep joint infections requiring re-operations
Badawy	Norway	1 year and 4 year follow up, average n/a	0	28,262	28,262	Skin incision to wound closure	<75, 75-90, 90-109, >110 minutes	n/a	Revision due to deep infection
Dicks	United States	Superficial infection: 30 days, deep infection: 1-year, average n/a	25,531	42,187	67,718	Skin incision to skin closure	Surgeon median operative duration: <79, 79-120, >120 minutes	NHSN SSI definition	NHSN deep SSI definition
Duchman	United States	30 days, n/a	40,275	59,169	99,444	Skin incision to skin closure	<60, 60-120, >120 mins (all based on CDC criteria); op times were further categorized into 30 min intervals	CDC SSI criteria	Any deep wound infection
Wang	China	1-year, n/a	9,481	7,861	17,342	Skin incision to skin closure	<60, 61-90, >90 mins	CDC SSI criteria	MSIS PJI criteria

Six unique analyses were performed: two utilized SSI as a primary endpoint, while four utilized PJI as the primary endpoint. Odds ratios for SSI and PJI were determined for operative time in each paired category.

Primary THA and TKA were evaluated in a pooled cohort and then analyzed separately as unique procedures. Analysis of THA alone with SSI as a primary endpoint could not be completed since enough studies with pertinent data, using our parameters, could not be found. Similarly, this was the case when evaluating TKA or THA alone using 90-minute cutoffs with PJI as a primary endpoint. Heterogeneity was tested using the Q and I2 statistics and could not be ruled out. As a result, inverse-variance weighted random-effects models were used to evaluate the pooled odds ratios using RStudio version 4.0.2 (RStudio Team (2020). RStudio: Integrated Development for R. RStudio, PBC, Boston, MA).

Results

A total of 81,373 THA and 149,319 TKA were evaluated from the six included studies, for a total of 230,692 primary TJA procedures. In a pooled cohort of THA and TKA, there was a 1.7 times higher likelihood of PJI in cases ≥ 90 minutes relative to those < 90 minutes in duration (OR=1.7 [95% CI 1.3-2.1]; p<0.001) [Table 3]. Using this cohort, there was also a 1.7 times higher likelihood of PJI in cases ≥120 minutes relative to cases < 120 minutes in duration (OR=1.7 [95% CI 1.0-2.7); p=0.048) [Table 4].

Table 3.

TJA PJI Results – 90 Minutes

Study	Author	Year	PJI ≥90	no PJI ≥90	PJI <90	no PJ <90
1	Badaway	2017	209	16135	102	11816
2	Wang	2019	32	2284	105	14921

OR: 1.65 (95%CI: 1.27 - 2.14); p-value: 0.0002

Table 4.

TJA PJI Results – 120 Minutes

Study	Author	Year	PJI ≥120	no PJI ≥120	PJI <120	no PJI <120
1	Aggarwal	2019	11	1261	39	4775
2	Anis	2018	42	2869	39	8890
3	Dicks	2015	94	7012	679	59933
4	Duchman	2016	52	17259	151	81982

OR: 1.63 (95%CI: 1.00 - 2.66); p-value: 0.0488

Using SSI as the primary endpoint in a pooled cohort of THA and TKA, there was 1.5 times greater odds of SSI in cases 120 minutes in duration versus cases < 120 minutes in duration (OR=1.5 [95% CI 1.2-1.8]; p<0.001) [Table 5]. In a cohort of TKA alone, there was also 1.5 times greater odds of SSI in primary TKA ≥ 120 minutes in duration versus primary TKA <120 minutes in duration (OR=1.5 [95% CI=1.1-2.1]; p=0.016) [Table 6]. Evaluating TKA alone, there was no increased odds of PJI in primary TKA ≥ 120 minutes in duration versus primary TKA < 120 minutes in duration (OR 2.0 [95% CI 0.8-5.3]; p=0.160) [Table 7]. Similarly, when evaluating THA alone, there was no increased odds of PJI in primary THA ≥ 120 minutes in duration versus primary THA < 120 minutes in duration (OR=1.1 [95% CI 0.8-1.4]; p=0.691) [Table 8].

Table 5.

TJA SSI Results

Study	Author	Year	SSI ≥120	no SSI ≥120	SSI <120	no SSI <120
2	Anis	2018	111	2800	199	8730
3	Dicks	2015	94	7012	679	59933
4	Duchman	2016	156	17155	479	81654

OR: 1.47 (95%CI: 1.18 - 1.83); p-value: 0.0005

Table 6.

TKA SSI Results

Study	Author	Year	SSI ≥120	no SSI ≥120	SSI <120	no SSI <120
1	Anis	2018	111	2800	199	8730
2	Dicks	2015	43	3402	390	38352

OR: 1.50 (95%CI: 1.08 - 2.08); p-value: 0.0159

Table 7.

TKA PJI Results

Study	Author	Year	PJI ≥120	no PJI ≥120	PJI <120	no PJI <120
1	Anis	2018	42	2869	39	8890
2	Dicks	2015	43	3402	390	38352

OR: 2.01 (95%CI: 0.76 - 5.30); p-value: 0.1566

Table 8.

THA PJI Results

Study	Author	Year	PJI ≥120	no PJI ≥120	PJI <120	no PJI <120
1	Aggarwal	2019	11	1261	39	4775
2	Dicks	2015	51	3610	289	21581

OR: 1.06 (95%CI: 0.80 - 1.39); p-value: 0.6909

Qualitatively, a study by Dicks et al.⁶ examined relationships between operative time and SSI, assessing THA and TKA separately. Operative time was recorded in two distinct methods: surgeon-median operative duration (the analysis we used for our meta-analysis due to increment categorization) and total operative duration. When analyzing 25,531 THAs, total operative durations exceeding the 75th percentile (>105 minutes) had a 1.11 times higher risk for SSI than durations in 25th-75th percentile (61-105 minutes) (RR 1.11; [95% CI, 1.03–1.21]; p=0.01).⁶ However, when analyzing surgeon-median operative duration in THAs, there was no significant increase in risk for SSI in cases >75th percentile (>104 minutes) in surgeon-median operative duration to cases in the 25th-75th percentile (71-104 minutes) (RR 1.10 [95% CI, 0.81– 1.50]; p=0.55).⁶ Similarly, Badawy et al.⁵ included 28,262 primary TKA procedures. There was not a significantly higher risk for postoperative infection when comparing the 90-110-minute group to the <75-minute group (Hazard Ratio (HR) 1.2 [95% CI 0.9-1.7]; p=0.30).⁵ However, there was a significantly higher risk of infection in the >110-minute group when compared to the <75-minute group (HR = 1.8 [95% CI 1.3-2.5], p=0.001).⁵

Discussion

The purpose of our meta-analysis was to assess the impact of increased operative time on subsequent infection rates in primary TJA. The results from our analyses indicate that for TJA procedures lasting greater than 120 minutes, there are significantly higher odds of PJI [Table 4] and SSI [Table 5] than in cases lasting less than 120 minutes. Similarly, operative time greater than 120 minutes in TKA procedures significantly affected SSI rates [Table 6], and TJA procedures exceeding 90 minutes were associated with higher PJI rates [Table 3].

Agodi et al.⁹ evaluated risk factors for infection in THA and TKA procedures. Study data demonstrated that longer operative times were a predictor of SSI in total joint procedures, similarly to the conclusions drawn from our study. This study also expanded their data to THA procedures, finding operative time as the single independent risk factor (RR: 4.54; 1.06-19.48).9 Unfortunately, we did not have enough data to quantitatively analyze THA SSI data independently.

In a study by Wang et al.,⁸ PJI and SSI rates were independently associated with increased operative time in TJA procedures. Data from this study was used in our meta-analysis on PJI rates, but the totality of literature review could not be accounted for due to low power in our study. They found that for every 20-minute increase in operative time there was a 25% increased risk for PJI.⁸ Also, there was a two times higher risk for developing SSI for patients exceeding an operation length of 90 minutes (OR 2.10 [95% CI 1.161-3.803], p=0.014).⁸ Duchman et al.⁷ ran a multivariate logistic regression using 99,444 patients to identify potential complications of increased operative times in TJA procedures. The notable variable studied for our purposes was any wound complication, including PJI. Operative time >120 minutes was found to be an independent risk factor for wound complication in TJA (OR 1.440, 95% CI 1.210-1.713).⁷ Additionally, they found that procedures >120 minutes had twice the incidence of wound complication than procedures ≤120 minutes.⁷

Both TJA [Table 5] and TKA [Table 6] procedures showed a significant effect of operative time greater than 120 minutes on SSI rate. As seen in Namba et al.,¹² procedure times taking less than 120 minutes were associated with lower rates of deep SSI compared to cases longer than 210 minutes. They also found that for every 15-minute increase in operative time there was a 9% higher deep surgical site infection risk (p<0.001; 95% CI: 0.04-0.13).¹² Both findings were in accordance with our quantitative analysis results.

In our subgroup analysis of TKA [Table 7] and THA [Table 8] there was no significant effect of increased operative time on PJI rate. Similarly, in a study by Aggarwal et al.³ there was no significant effect of operative duration greater than 120 minutes on PJI rates (p=0.840) following primary THA. In contrast to our findings, the study by Ong et al.¹³ found PJI rates to rise significantly with increasing operative time (p<0.001). This relationship showed a 78% increase in PJI risk for patients undergoing surgeries lasting longer than 210 minutes compared to those less than 120 minutes (OR 1.78, 95% CI: 1.40-2.26).¹³ However, Ong et al.¹³ did not include cases lasting 120-210 minutes in their analysis, which could significantly impact the study findings. A separate study by Ong et al.¹⁹ also studied the relationship between risk of revision in TKA and THA procedures and operative time. Analysis demonstrated increased rates of all-cause revision with increased operative time following both TKA and THA.¹⁹ Anis et al.⁴ supports the findings established by Ong et al.¹³ They found operative time to be an independent risk factor for both PJI (adjusted OR 1.01 [95% CI 1.01-1.012], p<0.001) and SSI (adjusted OR 1.01 [95% CI 1.004-1.01], p<0.001), with every 15-minute increase in time corresponding to higher infection rates.⁴ When using the 120-minute parameter, patients with surgeries greater than 121 minutes were 5 times more likely to develop PJI and 3 times more likely to develop SSI compared to procedures under 85 minutes.⁴

The low statistical power in our study may cause the discrepancies seen in our conclusion on THA and TKA compared to Anis et al.⁴ and Ong et al.¹³ The inclusion of more studies and patients could drive our data in the other direction, therefore aligning with conclusions set out by the previous studies. The difficulty associated with finding enough data to fit specific operative time increments is clearly a limitation in this portion of our quantitative analysis. Another limitation is that every meta-analysis is dependent upon the data in each study. All these studies are limited by their retrospective nature. Additionally, due to the variation in which data was reported we were unable to include a significant amount of existing data in the literature. The presence of confounding variables that could impact infection rates seen in cases of longer duration is likely, including obesity, overall case complexity (i.e., conversion arthroplasty), and intraoperative complications.

In conclusion, the rates of PJI and SSI in primary TJA procedures are significantly greater when exceeding a 120-minute threshold. Furthermore, a pooled cohort of THA and TKA demonstrated risk at a 90-minute threshold for PJI. Individual analysis of THA and TKA on SSI rates proved to be significant past the 120-minute point, however PJI results, due to limited data, were not significant. Findings from our quantitative results and qualitative review were relatively consistent, however additional analysis should be completed to compare specific operative time thresholds as more data with similar increment reports are made.