Incidence and risk factors for surgical site infection after medial opening-wedge high tibial osteotomy using a locking T-shape plate
Abstract
Medial opening-wedge high tibial osteotomy (MOWHTO) is a well-established surgical method for treatment of isolated medial compartment osteoarthritis with varus deformity, but the surgical outcomes may be compromised by surgical site infection (SSI). This study aimed to investigate the incidence and the risk factors for SSI after MOWHTO. This retrospective study included consecutive patients who underwent MOWHTO for isolated medial compartment osteoarthritis with varus deformity in two tertiary referral hospitals from January 2019 and June 2021. Patients who developed SSI within 12 months of surgery were identified by inquiring the medical records for index hospitalisation, notes of after-discharge outpatient visits, or records of readmission for treatment of SSI. Univariate comparisons were performed to detect the differences between SSI and non-SSI groups, and multivariate logistic regression analysis was used to identify the independent risk factors. Six hundred sixteen patients with 708 procedures were included and 30 (4.2%) cases of SSI occurred, with 0.6% rate for deep SSI and 3.6% for superficial. Univariate analyses showed significant difference between groups in terms of morbidity obesity (≥32 kg/m2) (20.0% vs 8.9%), comorbid diabetes (26.7% vs 11.1%), active smoking (20.0% vs 6.3%), time from admission to operation (5.2 ± 4.0 vs 4.1 ± 3.0), size of osteotomy ≥12 mm (40.0% vs 20.0%), type of bone grafting and lymphocyte count (2.1 ± 0.5 vs 1.9 ± 0.6). However, in the multivariate analysis, only active smoking (OR, 3.4; 95% CI, 1.4-10.2), size of osteotomy ≥12 mm (OR, 2.8; 95% CI, 1.3-5.9) and allogeneic/artificial vs no bone grafting (OR, 2.4; 95% CI, 1.0-10.8) remained significant. SSI was not uncommon after MOWHTO, but the majority was superficial. The identified three independent factors, including smoking, size of osteotomy ≥12 mm and allogeneic/artificial bone grafting would help risk assessment and stratification, target risk factor modification and clinical surveillance, and inform patient counselling.
Abbreviation
-
- ASA
-
- American Society of Anaesthesiologists
-
- BMI
-
- body mass index
-
- CI
-
- confidential interval
-
- COPD
-
- chronic obstructive pulmonary disease
-
- FBG
-
- fasting blood glucose
-
- IKDC
-
- International Knee Documentation Committee
-
- KOA
-
- knee osteoarthritis
-
- LCWHTO
-
- lateral closing-wedge high tibial osteotomy
-
- MOWHTO
-
- medial opening wedge high tibial osteotomy
-
- OR
-
- odds ratio
-
- RBC
-
- red blood cell
-
- SD
-
- standard deviation
-
- SSI
-
- surgical site infection
-
- WBC
-
- white blood cell
1 INTRODUCTION
During the past decade, medial opening-wedge high tibial osteotomy (MOWHTO) has gained more popularity for treatment of knee osteoarthritis (KOA), compared with the typical lateral closing-wedge high tibial osteotomy (LCWHTO), because it provides earlier rehabilitation, reduced frequency of neurovascular complications and ease to conversion to total knee arthroplasty.1, 2 However, the surgical results may be compromised by the higher incidence of complications, for example, patella baja, surgical site infection (SSI), delayed and nonunion, hardware failure and lateral cortex fracture.3-5
Among these complications, SSI is relatively uncommon but can be detrimental when it occurs, which may require multiple surgical revisions or even hardware removal, leading to undesirable functional results and even disability.6 Due to differences in study design, patient characteristics, surgical approaches, internal fixation devices, definition of infections and follow-up periods, the reported incidence rates of SSI varied greatly, from 0.39% to 4.9% for deep and 1% to 13.1% for superficial SSI.7-9 In 2013, Anagnostakos et al7 conducted a systemic review of infections after high tibial osteotomy, including 26 original studies of 2026 patients with 2210 knees, but did not specifically isolate MOWHTOs for analysis. Regarding the risk factors for SSI, inconclusive even contradictory results were reported due to various reasons regarding sample size, study population and others.7-11 Most recently, a Japanese national database including 12 853 HTO patients and identified artificial bone graft, longer anaesthesia duration, primary diagnosis of osteonecrosis and younger age as influential factors of SSI.8 However, in this study, the follow-up period (also namely the postoperative hospitalisation stay) was only a median of 34 days, which might have resulted in an underestimated SSI rate and the use of administrative data might introduce the bias in data collection.
A broad and deep understanding of the characteristics of SSI after MOWHTO would help surgeons in risk assessment and stratification and thus implement targeted precautious measures, as well as allow for improved patient counselling on complications. From a socioeconomic viewpoint, prevention is more cost-effective than treatment of SSI. Given that, this study focused on this cohort of patients undergoing MOWHTO to specifically address the incidence and risk factors for SSI.
2 METHODS
2.1 Data resource and inclusion and exclusion criteria
This retrospective study enrolled consecutive patients who underwent MOWHTO for treatment of isolated medial compartment osteoarthritis of the knee in the two tertiary referral and university-affiliated teaching hospitals (The First and The Third Hospital of Hebei Medical University) between January 2018 and July 2021. This study was conducted in accordance with the Declaration of Helsinki, and study protocol was approved by the ethics committee of the First Hospital of Hebei Medical University, which waived the requirement for informed consent due to the use of deidentified data for scientific research purpose.
The exclusion criteria were: age older than 70 years, a diagnosis other than primary OA, previous surgery on affected knee, use of fixation device other than locking plate, missing data on variables of interest or follow-up period <12 months.
2.2 MOWHTO procedure and postoperative management
In accordance with the method proposed by Miniaci et al,12 preoperative surgical planning was performed. General or neuraxial (epidural/spinal) anaesthesia was applied, based on patient condition and anaesthesiologist experience or preference. Prophylactic single-dose antibiotics (cefazolin, or vancomycin in cases of allergy) was administered 30 min prior to skin incision or tourniquet inflation. A longitudinal skin incision of about 6 cm was made anteromedially on the proximal tibia, 5 mm below the joint line skin and extending to the pes anserinus tendon, then was retracted distally; the distal superficial medial collateral ligament was released to expose the underlying periosteum. At the level of tibial tuberosity, the first guide K-wire was inserted from the anteromedial border of the tibia and advanced obliquely to the fibular tip. After confirmation by fluoroscope, the second guide wire was inserted posteriorly, parallel to the first wire, forming the primary osteotomy plane (parallel to the tibial slope). Osteotomy was performed in a biplanar manner, with the first parallel to the coronal plane, following by the second, namely the primary osteotomy. The opening wedge was formed by stepwise insertion of osteotomes, with anterior gap being approximately 2/3 of posterior gap. For cases with larger posterior opening-wedge, autogenous or allogeneic/artificial bone graft was filled in the medial osteotomy site to facilitate bone healing. A T-shape locking plate (Wego TM, Weihai, China; Double medical, Xiamen, China) was used to stabilise the osteotomy gap. At the discretion of the treating surgeon, a drainage tube was applied and typically removed on the postoperative day 2. Prophylactic antibiotics with 1 g of cefazolin Q8H for 1 day was administered in all patients.
Postoperatively, no brace or plaster cast was applied. Isometric quadriceps and active ankle exercises were initiated on the day one. Toe-touch mobilisation was permitted with assist of double crutches for the first 4 weeks, followed by partial weight-bearing assisted by one crutch at 6 weeks and full weight-bearing at 10 to 12 weeks post-operatively.
2.3 Definition of categorisation of SSI
SSI was defined according to the criteria issued by Centers of Disease Control and Prevention,13 which categorised it as superficial and deep. A superficial SSI involved only the skin and subcutaneous tissue within 30 days of surgery, on basis of the following signs and symptoms: erythema, swelling, heat and tenderness to palpation, which generally were resolved by empirical oral antibiotics. A deep SSI involved deep fascia or muscle within 12 months post operation, and was diagnosed if purulent discharge, abscess formation, wound dehiscence, plate fistula, fever and chillness were presented, which often required surgical debridement, intravenous antibiotics or even plate removement.
SSI cases were identified by reviewing the patients' medical records for the documented signs or symptoms regarding SSI and the records of microbial isolation and culture if any, during their index hospitalisation stay; after discharge, routine telephone visits at scheduled 1, 3, 6 and 12 months postoperatively were conducted to identify the SSI case, also the secondary hospitalisation or outpatient visit records were reviewed to determine whether a case of SSI was present.
2.4 Data collection
Two investigators (C.S. and G.J.) were responsible, but not independently, for data collection by inquiring the electronic medical and control quality, with disagreements resolved by a consensus via discussion. The relevant data were from five aspects: demographics (age, sex, living place and occupation before retirement, hospital stay), comorbidities (active smoking, alcohol drinking, body mass index [BMI], hypertension, diabetes, heart disease, cerebrovascular disease and chronic obstructive pulmonary disease [COPD], previous surgery), OA condition (affected side, preoperative radiographic Kellgren–Lawrence grade, preoperative IKDC [International Knee Documentation Committee] subjective score), surgery (time to surgery, anaesthesia mode, American Society of Anaesthesiologists [ASA] classification), surgical duration, size of osteotomy (in mm), intraoperative blood loss, bone grafting and laboratory testing after admission serum albumin concentration, red blood cell (RBC), haemoglobin, haematocrit, white blood cell (WBC), neutrophils, lymphocytes and fasting blood glucose (FBG).
2.5 Statistical analysis
Normality status of continuous variables was evaluated by the Shapiro–Wilkes test. Univariate comparisons were performed between patients with SSI (SSI group) and without SSI (non-SSI group), by Student-t test or Mann–Whitney U test for continuous variables based on their normality status, and by Chi-square or Fisher' exact test for categorical variables, as appropriate.
Multivariate logistic regression analysis using the stepwise backward method was performed, with variables tested with P < .10 in the univariate analyses entered, to identify the independent risk factors. Hosmer-Lemeshow test and the adjusted Nagelkerke R2 value were used to evaluate the goodness-of-fit of the final model, with P > .05 and R2 value of <.750 representing an acceptable fitting.14 Odds ratio (OR) with its 95% confidential interval (95% CI) were used to indicate the association magnitude. P < .05 as set as the statistical significance level and SPSS software 26.0 (IBM corporation, Armonk, NY) performed the all analyses.
3 RESULTS
There were 616 patients with 708 procedures (92 patients with bilateral procedures), including 138 males and 478 females, with an average of 57.3 ± 8.0 years. Twenty-seven orthopaedic surgeons performed all the procedures, with an average of 26.2 (range, 2-97) for one surgeon.
There were 30 operated knees in 30 patients who developed an SSI, representing an accumulated incidence rate of 4.2% (30/708). Among them, 4 (0.6%) were deep SSIs, which occurred at postoperative day 18, 47, 52 and 77, respectively, and were treated by arthroscopic or open debridement combined with intravenous antibiotics for 4 to 6 weeks, without removal of the hardware; at the subsequent follow-ups, bony union was reached in all four cases within 8 months. In the remaining 26 (3.6%) cases of superficial SSI, all were resolved with local wound care (multiple dressing change) and oral antibiotics for 1 to 2 weeks.
Univariate analyses showed significant difference between groups in terms of morbidity obesity (≥32 kg/m2) (20.0% vs 8.9%, P = .042), comorbid diabetes (26.7% vs 11.1%, P = .010), active smoking (20.0% vs 6.3%, P = .004), time from admission to operation (5.2 ± 4.0 vs 4.1 ± 3.0, P = .048), type of bone grafting and lymphocyte count (2.1 ± 0.5 vs 1.9 ± 0.6, P = .047). The multivariate analysis showed only active smoking (OR, 3.4; 95% CI, 1.4-10.2; P = .007), size of osteotomy ≥12 mm (OR, 2.8; 95% CI, 1.3-5.9; P = .008) and artificial vs no bone grafting (OR, 2.4; 95% CI, 1.0-10.8, P = .047) remained statistically significant. The goodness-of-fit of the final multivariate model was acceptable, with P = .471 for Hosmer-Lemeshow test and adjusted Nagelkerke R2 = .413.
4 DISCUSSION
Prevalent complications remain a clinically important issue after MOWHTO for treatment of osteoarthritis of isolated medial compartment of knee. In this study, we focused on the SSI after MOWHTO and found that the incidence rate was 4.2%, with superficial SSI accounting for the majority (26/30, 86.7%). We also identified three independent risk factors associated with SSI, including smoking, size of osteotomy ≥12 mm and allogenic/artificial bone grafting.
Our observed 4.2% rate of SSI after MOWHTO was within the range of reported rates (0%-18%) in literature,9, 15, 16 representing the median level. The greatly variable results reported in literature reflected the difference in sample size, population baseline characteristics and methodology, as well as whether SSI was studied as the primary result. Besides, the type of incision (L-shape, oblique or longitudinal) also contributed to the variation of incidence of SSI.1, 15, 17 Furthermore, Reischl et al11 identified oblique incision as the only risk factor for SSI. In fact, our finding of 4.2% of rate was consistent with a recent study specifically investigating SSI after MOWHTO that reported a rate of 3.6%.11 In most literature, SSI occurred being a most common complication or secondary to the lateral hinge fracture, but few studies focused on it likely due to the deep SSI accounting for only a small proportion with a very low incidence, as was reported by ours (four cases, 0.6% for incidence).
The role of smoking on complications following orthopaedic or other surgeries has been well-established. The underlying mechanisms may include its harmful effect on immune system (eg, suppressed immunoglobulin level, altered CD4 to CD8 ratio and reduced phagocyte activity), altering endothelial structure and function, and directly affecting wound healing process by nicotine, nitric oxide and carbon monoxide.18, 19 Additionally, smoking induced a greater risk of non-union after MOWHTO,20 and for this reason, some researchers even listed smoking as a relative contraindication in MOWHTO procedure.1, 21 Thus, it is suggested that smokers who intend to undergo MOWHTO should be emphatically informed of the particularly high risk of complications, especially the SSI and non-union; in our institution, for smokers, quitting smoking immediately after admission even at least 4 weeks before planned operation is mandatory to improve the pathophysiological conditions to reduce the risk of complications. The optimal timing of starting smoking cessation and the duration remain a subject focus in future researches (Tables 1 and 2).
| Variables | SSI group (n = 586) | Non-SSI group (n = 30) | P |
|---|---|---|---|
| Age (years) | 57.4 ± 7.8 | 54.8 ± 10.4 | .185 |
| <55 | 230 (39.2) | 16 (53.3) | .208 |
| 55-64 | 248 (42.3) | 8 (26.7) | |
| ≥65 | 108 (18.4) | 6 (20.0) | |
| Sex (males) | 134 (22.9) | 4 (13.3) | .222 |
| Occupation (manual work) | 411 (71.8) | 22 (73.3) | .709 |
| Living place (rural) | 430 (73.4) | 26 (83.3) | .226 |
| BMI (kg/m2) | 27.4 ± 3.5 | 27.4 ± 3.1 | .936 |
| Morbidity obesity (BMI ≥32 kg/m2) | 56 (21.5) | 6 (26.7) | .504 |
| Body weight (male, >90 kg; female, >75 kg) | 134 (22.9) | 4 (13.3) | .222 |
| Hypertension | 236 (40.3) | 12 (40.0) | .976 |
| Diabetes mellitus | 65 (11.1) | 8 (26.7) | .010 |
| Heart disease | 47 (8.0) | 2 (6.7) | .789 |
| Cerebrovascular disease | 27 (4.6) | 2 (6.7) | .604 |
| COPD | 18 (3.1) | 1 (3.3) | .936 |
| Previous surgery | 167 (28.5) | 9 (30.0) | .859 |
| Current smoking | 37 (6.3) | 6 (20.0) | .004 |
| Alcohol drinking | 118 (20.1) | 8 (25.0) | .506 |
| Operated side | .395 | ||
| Left | 278 (47.2) | 15 (50.0) | |
| Right | 218 (37.1) | 13 (43.3) | |
| Bilateral | 90 (15.7) | 2 (6.7) | |
| Preoperative IKDC | 43.7 ± 6.3 | 44.1 ± 8.1 | .816 |
| Kellgren and Lawrence grade | .182 | ||
| 1 | 67 (11.4) | 1 (3.3) | |
| 2 | 237 (40.4) | 10 (33.3) | |
| ≥3 | 282 (48.2) | 19 (63.4) | |
| Time from admission to operation (days) | 4.1 ± 3.0 | 5.2 ± 4.0 | .048 |
| ≤3 | 268 (45.7) | 13 (43.3) | .797 |
| >3 | 318 (54.3) | 17 (56.7) | |
| Total hospitalisation stay (day) | 11.8 ± 4.3 | 12.1 ± 4.7 | .755 |
| ASA classification | .341 | ||
| I-II | 466 (79.5) | 26 (86.7) | |
| III-IV | 120 (20.5) | 4 (13.3) | |
| Anaesthesia mode | .397 | ||
| Regional | 261 (44.5) | 11 (36.7) | |
| General | 325 (55.5) | 19 (63.3) | |
| Surgical duration (minutes) | 121.7 ± 60.5 | 120.3 ± 58.0 | .903 |
| ≥180 | 91 (15.5) | 6 (20.0) | .512 |
| Intraoperative blood loss (ml) | 130.9 ± 112.3 | 133.8 ± 124.0 | .893 |
| ≥300 | 65 (11.1) | 2 (6.7) | .448 |
| Bong grafting | .042 | ||
| No | 129 (22.0) | 3 (10.0) | |
| Autologous | 103 (17.6) | 2 (6.7) | |
| Allogeneic/artificial | 354 (60.4) | 25 (83.3) | |
| Size of osteotomy (≥12 mm) | 117 (20.0) | 12 (40.0) | .009 |
| Albumin level (g/L) | 42.2 ± 3.5 | 41.9 ± 3.1 | .576 |
| FBG | 5.4 ± 1.3 | 5.6 ± 2.8 | .373 |
| WBC (*109 /L) | 6.5 ± 2.1 | 6.3 ± 1.3 | .618 |
| Neutrophil count (*109/L) | 4.0 ± 2.0 | 3.6 ± 1.1 | .353 |
| Lymphocyte count (*109/L) | 1.9 ± 0.6 | 2.1 ± 0.5 | .047 |
| RBC (*1012/L) | 4.3 ± 0.5 | 4.2 ± 0.4 | .541 |
| Haemoglobin (g/L) | 129.8 ± 14.8 | 129.0 ± 11.1 | .754 |
| Haematocrit (%) | 38.9 ± 4.2 | 38.6 ± 3.3 | .700 |
- Abbreviations: ASA, American Society of Anesthesiologists; BMI, body mass index; COPD, chronic obstructive pulmonary disease; FBG, fasting blood glucose; IKDC, International Knee Documentation Committee; RBC, red blood cell; WBC, white blood cell.
| Variables | OR and 95%CI | P |
|---|---|---|
| Active smoking (vs no) | 3.4 (1.4-10.2) | .007 |
| Size of correction (≥12 mm) | 2.8 (1.3-5.9) | .008 |
| Bone grafting | ||
| No | Reference | |
| Autologous | 1.2 (0.2-5.6) | .845 |
| Allogeneic/artificial | 2.4 (1.0-10.8) | .047 |
The positive correlation between the size of osteotomy and the incidence of complications has been reported by several studies, consistent with ours (OR, 2.8; P < .05), but the cut-off value might be of somewhat difference. In studies by Nelissen et al,22 authors found the significantly more complications in patients with osteotomy size >10 vs ≤10 mm (60% vs 34.5%, P = .05), and over half (53.3%) of these complications were attributed to intrinsic instability at the osteotomy site. Most recently, isolating SSI from the overall complications, Zhao et al10 found 1.7-fold increased risk of SSI associated with an osteotomy size >10 mm over ≤10 mm in a retrospective of 114 MOWHTO procedures. The larger osteotomy gap meant the more surgical trauma to bone and the surrounding soft tissue and the poorer stability at the osteotomy site, possibly, resulting in an increased risk of wound infection. Another viewpoint showed that the larger osteotomy gap increased the probability of lateral hinge fracture, thereby inducing the increased risk of infection,23 but the very low proportion (2.2%) of lateral hinge fracture in our study is hardly likely to generate a significant result.
Bone grafting, especially the type (autologous, allogeneic, artificial), was frequently discussed together with the gap of osteotomy because of their intrinsic correlation. In this study, compared with no bone grafting, allogeneic/artificial bone grafting was associated 2.4-fold increased risk of SSI, but not was the autologous bone grafting. Indeed, bone grafting has dual effects: facilitating the bone union of osteotomy gap and stabilisation of the gap during healing period especially for cases with a gap ≥10 mm,24 and increasing the risk of weak antibacterial property, bio-incompatibility and immunological rejection with allogeneic/artificial bone graft.25 This could largely explain the greater risk of SSI when allogeneic/artificial bone graft was used. Despite avoiding the risks associated with allogeneic/artificial bone grafting, autogenous bone grafting was not without risk, for example, the more surgical trauma, longer time for the overall operative procedure and the donor site morbidity,26, 27 which might be another cause for the SSI for osteotomy site. In a recent study, Kim et al28 combined allograft bone with autologous bone marrow for osteotomy gap filling after MOWHTO, exhibiting a benefit on osteoconductivity but not on clinical results over allograft bone alone. This is an interesting attempt and future research is warranted to provide more clinical and radiographic evidences.
The strengths of this study included a relatively large sample in two university affiliated hospitals, and inclusion of multiple potential factors for adjustment, however, several limitations must be noted. First, the retrospective study had inherent limitation in sample selection and data collection. Furthermore, patients' self-reported comorbidities may not represent the true conditions, likely because of asymptomatic conditions (eg, hypertension) existing in a certain proportion of population. Second, although we identify the smoking as an independent factor for SSI in this study, the detailed data on quantifying the smoking (number of cigarettes smoked per day, and the smoking years) are not captured; indeed, this effect might be dose dependent because of accumulated harm of smoking on tissues. Third, cases of SSI were determined by inquiring the medical records for index hospitalisation, outpatient visits or for readmission, and thus possibly that patients with SSI were not readmitted or treated elsewhere, which might have underestimated the incident SSI. Fourth, there are still unmeasured patient-, surgeon- or nurse-level confounders, for example, medications use, experience and wound care quality that influenced the risk of SSI, thus leaving the residual confounding. Fifth, despite a two-institution design, both of them were university affiliated, thus more likely admitting patients with severer knee condition or poorer pathophysiological status. Thus, our results were generalisable to other lower-level medical settings.
In conclusion, we found the 4.2% rate of SSI after MOWHTO for treatment of isolated medial compartment osteoarthritis of the knee and identified smoking, size of osteotomy ≥12 mm and allogeneic/artificial bone grafting as independent risk factors. These findings would help risk assessment and stratification, target risk factor modification and clinical surveillance, and inform patient counselling.
AUTHOR CONTRIBUTIONS
Xiaoguang Yu conceived and designed the study; Guoxing Jia, Jin Xie and Jia Li inquired the medical records and collected the relevant data; Wei Dong performed the analyses and interpreted the results; Sen Liu prepared the tables and figures; Guoxing Jia drafted the manuscript and all authors (Guoxing Jia, Congcong Sun, Jin Xie, Jia Li, Sen Liu, Wei Dong, Xiaoguang Yu) reviewed and approved for publication.
ACKNOWLEDGEMENTS
We are grateful to G.L. of department of Orthpaedics Surgery and X.Z. of Department of Epidemiology and Statistics for their kind help.
FUNDING INFORMATION
This study was supported by the Hebei Province Medical Science Research Project (No. 20211310).
CONFLICT OF INTEREST
All of the authors declared no conflict of interest.
ETHICS STATEMENT
The study protocol was approved by the ethics committee of the First Hospital of Hebei Medical University.
Open Research
DATA AVAILABILITY STATEMENT
In accordance with institutional policy for patients' medical records, data used in this study are not open to public, but will be available upon motivated request to the corresponding author for purpose of scientific research.
