Introduction

Arthroscopy is a valuable tool for treating mechanical symptoms related to early stages of knee arthritis with satisfactory postoperative clinical results. According to Pagnano et al. [14], Miller et al. [11] and McGinley et al. [10], arthroscopic procedures can improve clinical symptoms and delay total knee arthroplasty (TKA) intervention, even in middle-aged and geriatric populations [21].

At later stages of arthritis, total knee arthroplasty is presented as a predictable and reproducible surgical procedure having an estimated mean implant life of 10 years in 92% of the cases. World literature cites the clinical efficacy of this surgical intervention in improving adverse symptoms of arthritis with even total elimination of pain and better range of motion, promoting independence and improved quality of life [2, 5, 20].

However, it must be underscored that previous knee surgery is associated with surgical difficulties and postoperative complications in the case of primary TKA on the same knee [4]. Previous surgeries such as tibial or femoral osteotomy, distal patellar–femoral re-alignment, osteosynthesis of tibial plateau or distal femoral fractures could result in poor postoperative clinical scores in primary TKA as compared to cases without previous knee surgery.

In the literature, the role of arthroscopy as a previous surgical procedure in primary TKA has not yet been established. Arthroscopy is a minimal invasive surgery performed for minor surgical trauma, reduced morbidity and shortens hospitalization period when compared with other surgical interventions. Hence, this type of surgery before TKA could be considered a minor surgical procedure with minimum postoperative complications.

Therefore, is previous knee arthroscopy related to worse results in primary TKA?

Materials and methods

From March 1988 to August 2005, 1,474 primary TKA were performed in our department. The posterior-stabilized tricompartmental TKA with one-third medial condyle (Tornier®, Saint Ismier, France) was used.

A retrospective and cohort study analyzed two distinct groups of patients. In one group of 1,119 primary TKA (1,035 patients) (group A) had no previous knee surgery, while 60 primary TKA (60 patients) (group B) had undergone previous arthroscopy. Tables 1, 2, 3 present the percentage distribution of patients according to age, sex, side, and body mass index (BMI), the etiological diagnosis and type of knee arthritis in groups A and B, respectively. The interval between arthroscopy surgery and primary TKA was variable, ranged from a minimum interval period of 4 months and a maximum of 120 months (mean interval of 53 months).

Table 1 Percentage distribution of patients according to age, sex, side, weight, height and BMI in group A (1,035 patients) and group B (60 patients)
Full size table
Table 2 Percentage values of primary TKA according to the etiological diagnosis in groups A and B
Full size table
Table 3 Percentage values of primary TKA according to type of knee arthritis in groups A and B
Full size table

All the patients underwent a prior systematic assessment that included a clinical examination, radiographs [bipodal hip–knee–ankle film (pangonogram) [12], anteroposterior knee view with monopodal support, posteroanterior view at 45° knee flexion, lateral view at 30° knee flexion, patellar axial at 30° knee flexion, stress valgus and varus X-ray using a Telos® system] and International Knee Society (IKS) knee and functional scores [7]. Blackburne–Peel index (BPI) [3] was used as radiological measure for determining pre and postoperative patellar height. Pre and postoperative patella position in the radiographic axial view at 30° knee flexion were evaluated.

In our series of cases, the clinical and radiological data for each patient were obtained at the latest follow-up before any revision TKA.

In group A, the medial approach was used in 974 cases (87%) involving 10 cases of tibial tubercle osteotomy, the lateral approach in 134 cases (12%) involving 64 cases of tibial tubercle osteotomy and anterolateral approach was used in 11 cases (1%).

However, in group B, the medial approach was used in 53 cases (88.3%) involving 1 case of tibial tubercle osteotomy, the lateral approach was used in 7 cases (11.7%) involving 4 cases of tibial tubercle osteotomy.

The mean duration of surgery was 86.19 and 105.38 min for groups A and B, respectively. No difference was found for both study groups (P = 0.34).

Statistical analysis

In our primary TKA series, the statistical analysis of certain variables showed that these two study groups of patients were similar (regarding some characteristics), and the ANOVA confirmed that these groups could be comparable (no statistical difference for height, weight, BMI and sex distribution). The data obtained underwent statistical analysis with Minitab® Statistical Software Package (Minitab Ltd. State, Brandon Court, Unit E1, State College, Pennsylvania). In this statistical analysis, the limit of significance considered was 5%.

Results

During the postoperative follow-up, 86 patients (8%) were lost from the original group A and 949 patients (992 primary TKA) were re-assessed regularly. In group B one patient (2%) was lost and 59 cases (98%) re-assessed. The mean follow-up period was 43.2 months (24–193) and 46.6 months (24–140), respectively for groups A and B.

Knee range of motion

When comparing groups A and B, the preoperative mean knee range of motion data for hyperextension, lag of extension and hyperflexion were respectively, 0.4°/4°/116° and 0.5°/5°/115° (P = 0.76, P = 0.4 and P = 0.47, respectively). Likewise, in the postoperative evaluation, the mean knee range of motion was statistically similar for both study groups: 0.7°/0.5°/118° and 0.9°/0.2°/117° respectively, in the groups A and B (P = 0.5, P = 0.5 and P = 0.5, respectively).

IKS knee and functional scores

The pre and postoperative IKS knee score did not differ for both A and B study groups (P = 0.6 and P = 0.23, respectively). In the same way, pre and postoperative IKS functional score showed no difference between both study groups (P = 0.11 and P = 0.10, respectively). The mean IKS knee and functional scores are presented in Table 4.

Table 4 Pre and postoperative mean IKS knee and functional scores in both study groups
Full size table

Level of pain

The statistical analysis of preoperative evaluation of both study groups showed no difference. However, the postoperative evaluation of both study groups showed that the distribution of patients classified as no pain and severe level of pain differed (P < 0.03 and P < 0.003, respectively) (Table 5).

Table 5 Percentage values of pre and postoperative of level of pain between both study groups
Full size table

Radiological evaluation

In group A, arthritis was radiologically classified according to Ahlbäck’s classification [1], in 1% of the cases as stage 1, 24%—stage 2, 47%—stage 3 and 28% of the cases as stage 4. In group B, the cases were classified as stages 1–4 in 6.5, 33, 54 and 6.5%, respectively.

Statistical analysis showed a difference in distribution of cases classified as stages 1 and 4 in groups A and B (P = 0.01 and P = 0.01, respectively). On the other hand, there was no difference between cases classified as stages 2 and 3 in both study groups.

The pangonogram radiograph measured knee malalignment and the percentage values were grouped under four intervals: 0°–3°; 4°–6°; 7°–9° and above 9° are presented in Table 6.

Table 6 Distribution of knee malalignment at four levels of pre and post pangonogram in both study groups
Full size table

Patella position in the radiographic axial view

The percentage values of pre and postoperative of patella position in the radiographic axial view at 30° knee flexion and statistical analysis are presented in Table 7.

Table 7 Percentage values of pre and postoperative patella position in the radiographic axial view at 30° of the knee flexion in both study groups
Full size table

Patella height

The pre and postoperative mean values of patella height measured by Blackburne–Peel index [3] in the lateral radiographic at 30° knee flexion were respectively 0.76 and 0.60 for group A; 0.75 and 0.59 for group B. Statistical analysis showed no difference in both study groups in the pre and postoperative values (P = 0.5 and P = 0.48, respectively).

Intraoperative complications

Tibial plateau or femur fissures (minor and non-displaced fracture) and collateral ligaments damage (partial tear, never a complete disruption) were the most frequent intraoperative complications in primary TKA surgery. Intraoperative complications were reported in 44 cases of group A (4%) and 4 cases of group B (7%). However, statistical analysis showed no difference between both study groups (P = 0.3).

Postoperative complications

At the mean postoperative follow-up of 43.2 months, we identified 103 complications in group A (10% of the cases in this group). A second surgical intervention was performed in 80 cases (79%) and prosthetic component replacement was performed in 33 cases (41% of these cases).

At the mean postoperative follow-up of 46.6 months, we registered 18 complications in group B (30% of the cases in this group). 100% of these cases underwent another surgical intervention and in 39% of these cases (7 cases), the prosthetic components were replaced.

Regarding the same type of postoperative complications (Table 8), we observed that group B had a higher rate of postoperative complications and percentage of prosthetic components replacement when compared to group A (P < 0.001 and P < 0.001).

Table 8 Distribution of cases according to type of local complication in the series of primary TKA in groups A and B
Full size table

Survival rates

At 120 months follow-up, the Kaplan–Meier curve showed a 98.1% survival rate (95% confidence interval, 99.1–97.1%) for group A, while group B had 86.8% survival rate (95% confidence interval, 99.5–74.1%).

Discussion

Scher et al. [17] and Jerosch et al. [8] emphasize that arthroscopic management is an effective method proposed for the treatment of some postoperative TKA complications, such as knee stiffness, painful TKA (patellar clunk syndrome and painful patellofemoral crepitus), intraarticular loose body and pseudomeniscus. Similarly, Klinger et al. [9] states that arthroscopy could be useful in relieving pain, improving knee function and even, in elucidating undiagnosed cases of painful knee arthroplasty.

Although knee arthroscopy is considered a minimal invasive surgery on a native knee, it is not free from postoperative complications. Sherman et al. [19] in a retrospective review of 2,640 patients reported 8.2% postoperative complications, of which 58% were classified as major complications such as infections, hemarthrosis, adhesions, instrument breakage, cardiovascular and neurological problems.

On the other hand, previous knee surgery seems to be related to inferior clinical and functional postoperative results as well as postoperative primary TKA complications. Nelson et al. [13] concluded that TKA after varus osteotomy of the distal femur is a technically demanding procedure and associated with inferior clinical results when compared with those of primary TKA without previous osteotomy. Similarly, Parvizi et al. [16] reported that primary TKA after high tibial osteotomy presented a high rate of radiographic evidence of TKA loosening. Analyzing primary TKA in patients with previous distal femoral fracture, Papadopoulos et al. [15] observed that these cases resulted in restricted range of motion and intraoperative complications.

But in fact, the influence of arthroscopy as previous surgery in the outcome of primary TKA has not been reported.

In our work, the groups A and B presented no different characteristics regarding sex, side, BMI and type of knee arthritis (Tables 1, 3). Similarly, in both study groups, the main etiological diagnoses were arthritis, followed by inflammatory rheumatism. Moreover, preoperative evaluation IKS knee and functional scores and knee range of motion did not differ. The mean age of group B was 3 years less than group A (P < 0.001). In relation to the stage of arthritis, 75% of cases in group A and 60% of group B were classified as stages 3 and 4. Statistical analysis showed a prevalence of stage 4 cases in group A.

In our series, 52 and 57% of cases presented knee malalignment superior to 7º in groups A and B, respectively (P = 0.35). However, postoperative evaluation showed that surgery improved the knee alignment and no difference was found in both groups (Table 6).

As shown by some authors [6, 18], intraoperative complications during primary TKA (tibia or femur fracture, disruption of collateral ligament or even patellar tendon disruption) could influence postoperative outcome. However, in our series of cases, this occurrence was not identified in group B (P = 0.3).

Improved preoperative of IKS knee and functional scores were 107 and 43% for group A and 96 and 45% for group B. The statistical analysis showed no difference in improved IKS knee and functional scores (P = 0.2 and P = 0.4, respectively).

Despite the fact that there was no statistical difference in IKS knee and functional scores in our series, group B presented a higher rate of postoperative complications when compared to group A (P < 0.001) (Table 8 presents patella baja, patella fracture, anterior tibial tubercle fracture, TKA loosening and stiffness as postoperative complications in group B). This was demonstrated by the fact that 30% of the cases were re-operated in group B and 8.3% of these cases were submitted to revision TKA, while only 4% of cases in group A were re-operated and 1.4% underwent surgical revision (P < 0.001 and P < 0.001, respectively). These results are also confirmed by the Kaplan–Meier survival curves, with a survival rate of 98.1 and 86.8% at 10 years follow-up for groups A and B, respectively.

In group B, the mean interval between arthroscopy and primary TKA was 53 months. In this group, we observed that in 57% of the cases the mean interval was 25.6 months. However, in this study, statistical analysis did not demonstrate a direct correlation between arthroscopy and primary TKA interval and postoperative complications and failures (P = 0.55). Although the mechanism involved remains to be clarified, we hypothesized that knees with more inflammatory soft tissue (group B) could be related to higher postoperative complications and failure rates.

Even though, our results did not identify the mechanism by which previous knee arthroscopy influences primary TKA outcome, these data have motivated us to follow-up this study and conduct further extensive studies on patients who have undergone primary TKA after other knee surgeries on the same knee.

Conclusion

Our data allow us to conclude that previous knee arthroscopy should be considered a factor related to postoperative primary TKA outcomes as demonstrated by the higher rate of postoperative complications and failures (P < 0.001) as well as a worse survival curve than group A.