1 Introduction

Patellofemoral instability (PFI) is a relatively common knee pathology, particularly affecting adolescents and young athletes, occurring mostly in patients aged 10–19 years, with a higher prevalence in females and those participating in contact sports.¹ The recurrent episodes of dislocation can lead to functional impairment and repetitive chondral damage, and, eventually, early patellofemoral (PF) osteoarthritis.²

Management of patellofemoral instability is challenging due to its multifactorial pathogenesis, which involves both osseous and soft tissue abnormalities. Key anatomical factors include medial patellofemoral ligament (MPFL) deficiency, vastus medialis obliquus hypotrophy, trochlear dysplasia, increased tibial tubercle–trochlear groove (TT-TG) distance, patella alta, and excessive patellar tilt.^3,4 Treatment remains challenging, and the best therapeutic approach is yet to be settled.⁵ Recurrence rates can be as high as 50 %, especially in maltracking abnormalities, such asTT-TG superior to20 mm. Tibial tubercle osteotomy (TTO) has settled as a cornerstone surgical intervention for addressing this biomechanical malalignment. Long-term studies reveal variable results, with 62.5 % of patients maintaining good to excellent results after 10–15 years of the surgery, with the outcomes correlated strongly with pre-existing chondral damage severity.⁶ Recent systematic reviews demonstrate consistent improvements in pain scores and functional metrics, yet 16–21 % require reoperations, predominantly for hardware removal.⁷ Recently, medial patellofemoral ligament reconstruction (MPFLR) has emerged as a valid option, for PFI, whether isolated or combined with TTO. Controversy persists regarding optimal indications, as the combined procedure shows lower revision rates versus isolated reconstruction, while anteromedialization techniques demonstrate 60–93 % satisfaction rates in maltracking correction.^8–10 On the other hand, there are some studies reporting PF joint degeneration as a consequence of the increased contact pressures, ultimately producing late osteoarthritis.^11,12 Thus, these divergent outcomes underscore the critical need for individualized surgical planning based on anatomic variants and articular cartilage status.

The available literature shows that most studies focus on the short-term results, often 5 years or less,⁷ existing a lack of evidence concerning the long-term outcomes. To address this, we undertook a comprehensive investigation, with a minimum 10-year follow-up, using qualitative magnetic resonance imaging (MRI) analyses and clinical evaluations to assess our long-term results, this beingour primary goal. Additionally, we aimed to identify risk factors that could predict suboptimal clinical outcomes.

2 Material and methods

We designed a retrospective study, identifying all patients submitted to tibial tubercle osteotomy between 2008 and 2013 in a tertiary hospital. The primary indication for surgery was recurrent lateral patellar instability, with at least one episode of patellofemoral dislocation, and a TT-TG distance superior to 15 mm. Patients with concomitant knee pathology, like meniscal or ligamentar lesions, were excluded. The minimal follow-up time for enrollment was 10 years.

This project was submitted and accepted by our ethical committee (number 242–22; Local Health Unit of São João, Porto, Portugal).

All surgeries were executed by senior surgeons who belong to a knee-specific unit, using the same standard technique. The procedure started with a longitudinal incision lateral to the patellar tendon, allowing access to the lateral retinaculum, which was released to reduce lateral patellar tension; afterwards, the tibial tubercle osteotomy was prepared, with a distally based periosteal hinge to maintain the blood supply, with a length of 5–6 cm. The fragment was, then, medialized to achieve a TT-TG distance, ideally, inferior to 10–12 mm. Range of motion (ROM), patellar tracking, and patella glide were observed at the end of the surgery and, after that, the bone plug was fixed with 1–2 cancellous lag screws. Concerning weight-bearing protocol, the decision was not always consensual, varying between no restrictions in the postoperative period and walking with crutches for four to six weeks, with partial weight-bearing.

After the minimum follow-up period of 10 years, patients were evaluated. Data collected included demographic and clinical information such as age, gender, concurrent surgical procedures, and complications like recurrent dislocations, episodes of subluxation, or loss of ROM. The incidence of revision surgeries was also documented. Functional outcomes were assessed using the Kujala Anterior Knee Pain Scale and the Tegner Lysholm Score.

Radiological analyses employed different imaging modalities pre and postoperatively. Preoperative PF relations were exclusively obtained through computed tomography (CT) scans. To reduce radiation exposure, postoperative relations of patellofemoral anatomy were conducted using magnetic resonance imaging (MRI). These scans were performed on 1.5T systems (GE Healthcare and Siemens Healthcare) using a standardized protocol, including multiplanar T1 and T2-weighted sequences without contrast. Both imaging acquisitions were performed according to the same institutional protocol, with the knee maintained in full extension (0°) under passive and quadriceps-activated conditions. Multiple studies have validated the reliability of MRI for obtaining these relations, demonstrating strong agreement with CT-based evaluations.^13,14 Images were reviewed using Sectra UniView Picture Archiving and Communication System workstations. Two reviewers conducted the radiological measurements, with oversight and guidance from specialized musculoskeletal radiologists to ensure accuracy.

Regarding patellofemoral measures, patellar alignment, patellar height, patellar tilt, and patellar maltracking (TT-TG distance) were considered. Patterns of malalignment were defined and divided into four types, according to the classification proposed by Fulkerson.¹⁵ Patella height was determined via the Insall-Salvati index through X-ray.¹⁶ Lateral patellar tilt was evaluated by calculating the angle between a reference line across the patella and a tangential line along the posterior femoral condyles, seen in axial MR images, with and without quadriceps contraction¹⁷. Lateral patellar displacement was calculated using an axial slice: a line between the medial and lateral trochlear facet was drawn, then a perpendicular line at the trochlear sulcus and the patellar spine level, and the distance between the 2 intersections was measured¹⁸ (Fig. 1).

Close

Fig. 1 – Patellar displacement measurement on axial MRI images, by first drawing a line connecting the medial and lateral trochlear facets. A perpendicular line was then drawn at the level of the trochlear sulcus and the patellar spine. The distance between the intersection points of these lines was recorded as the patellar displacement. The left image represents the knee in full extension (0°), while the right image demonstrates the measure during quadriceps activation.

Export to PPT

Trochlear dysplasia parameters included sulcus angle and lateral trochlear inclination, were calculated in an axial slice.¹⁷ Patellofemoral and tibiofemoral chondropathy were evaluated postoperatively, through the MR images, using the modified Outerbridge classification¹⁹ (Fig. 2).

Close

Fig. 2 – Axial slice of a T2-weighted sequence showing grade IV chondropathy, classified according to Outerbridge, mainly on the lateral facet of the patella.

Export to PPT

Data was collected and stored, anonymized, using Microsoft Excel® (Version 2016; Microsoft Corp), with statistical analysis conductedwith IBM SPSS® 30. For continuous and categorical variables, descriptive statistics were calculated. Continuous variables were reported as weighted mean and estimated standard deviation or median and interquartile range (IQR), based on normal distribution, whereas categorical variables were reported as frequencies with percentages.

Categorical variables were processed using Fisher's exact or Chi-square test, depending on the count of the cells. The independent or paired t-test for normally distributed variables and the nonparametric Mann–Whitney U test, or Wilcoxon signed-rank test, were used to compare categorical and continuous variables. Continuous variables were assessed through the Spearman or Pearson test. A value of p < 0.05 was considered statistically significant.

3 Results

50 patients met the inclusion criteria. The final sample comprised 27 individuals, 36 knees, due to challenges in contact, with loss of follow-up. The median follow-up time was 12.5 years. The initial mean age was 19 IQR 17–34 years, characterized by a group of mostly women (65.7 %).

We did not have any immediate or short-term complications to report, such as infection or fractures. All patients achieved full ROM after the first month. We did not have any cases of non-union. Symptomatic hardware was the most frequent complications with screw removal in half of the patients. We had 5 cases of recurrent patellar dislocation, all occurring in a later postoperative period rather than immediately following surgery. This subgroup comprised 80 % women, with a median age of 16. The episodes were related to a new trauma event in patients who persisted with some degree of abnormal anatomy, like high TT-GT distance, tilt, or persistent subluxation.

We had one case of patella baja that had its osteotomy revised, with resolution of pain and instability. Trochleoplasty was an associated procedure in two patients with severe trochlear dysplasia (>165°), with reduction of the sulcus angle to normal values. Vastus medialis obliquus (VMO) advancement was performed in 16.6 % of the cases, as a part of the combined osteotomy and lateral release approach, when medial soft tissue insufficiency was found.

Table 1 resumes this data.

The median and the interquartile range for patellar instability relations are available in Table 2.

Patellar height remained roughly the same between pre and postoperative periods (1.2 IQR 1–1.38 V S 1.2 IQR 1–1.5, respectively). The same can be affirmed for the measures of trochlear dysplasia: sulcus angle (136 IQR 127-140° VS 143 IQR 135-147.5°) and lateral trochlear inclination (13 IQR 12-18° VS 13 IQR 10.5–15.5°).

We observed a significant decrease in the patellar tilt and patellar displacement between the preoperative and postoperative periods. Both values were reduced by half when measured without quadriceps contraction (patellar tilt: 28.5 IQR 17.75–34.25 V S 16.5 IQR 10–24.25°, p < 0.05; patellar displacement: 6.3 IQR 3.07–14.6 V S 2.5 IQR 0–8.25 mm, p = 0.01). When the quadriceps was activated, a significant decrease of the patellar tilt was also observed (34 IQR 24–42.5 V S 29 IQR 15-36° p = 0.014). About patellar displacement, although an improvement was observed, the reduction was not statistically significant (10.7 IQR 3.8–15 V S 7 IQR 2–13 mm, p = 0.191). Regarding malalignment, the same significance was observed for the TT-TG distance, with an important diminution of this parameter (23 IQR 15.05–25.3 V S 15 IQR 10–18 mm, p = 0.002). Around 60 % of the patients presented with a type II pattern of malalignment (subluxated and tilted).

Concerning patello-femoral chondral lesions, 54.8 % were classified in types III and IV of the modified Outerbridge classification. Chondral lesions demonstrated a homogeneous distribution: 23.3 % localized to the lateral patellar facet, 13.3 % to the patellar apex, 10 % exhibiting diffuse involvement, 10 % on the medial patellar facet, and 6.7 % isolated to the trochlea alone. Despite no correlation (p = 0.439), most patients had a subluxated and tilted patella (pattern II). Approximately 35.5 % of patients did not develop chondral alterations, with a median age at the time of admission of 18 years. No correlation was found between their presence and actual age (p = 0.092), gender (p = 0.661), TT-TG distance (p = 0.546), patellar height (p = 0.059), sulcus angle (p = 0.756), patellar tilt (p = 0.374), or displacement (p = 0.257).

Nonetheless, the patients that had a higher grade of chondral involvement were older (25 IQR 17.5–37 V S 18 IQR 17–20.25), had a slightly higher TT-GT distance (20.3 IQR 15.15–22.2 V S 18.25 IQR 16.45–21.65), and a higher patella as well (1.3 IQR 1.22–1.49 V S 1.05 IQR 1–1.35).

80 % of the patients had no osteoarthritis in the tibiofemoral (TF) compartment. Concerning the remaining 20 %, 7.4 % corresponded to grades II and IV of chondropathy. When present, they were mainly in the medial compartment. No correlation was found between its presence and age on admission (p > 0.05).

The median Kujala score was 89.5 IQR 82.5–97.5, and the median Lysholm was 91 IQR 81.75–95.75. A subgroup analysis was performed based on the TT-TG distance, patellar tilt, displacement, and PF and TF chondropathy. We compared them to ascertain if any of these factors could have a potential functional impact, but we did not find any relevant clinical significance (p > 0.05). However, we did see a tendency showing that patients with no malalignment (pattern IV) evolved with better functional outcomes over time (Kujala and Lysholm superior to 95).

4 Discussion

The main findings of our paper are the significant and consistent decrease of the TT-GT distance, patellar tilt, and patellar displacement, after 10–15 years of TTO (p < 0.05). Furthermore, we also found that around 50 % of the patients evolved with grade III and IV chondral lesions, and around 20 % had some chondropathy on the TF joint.

The approach and surgical management of PFI involves essentially three established procedures: MPFL reconstruction to restore medial restraint, TTO to realign the extensor, and trochleoplasty to address underlying trochlear dysplasia. Among these, TTO has been used for decades, remaining a very popular procedure.^19–21

The rationale behind TTO is to reduce the lateral force vector exerted on the patella through its medialization, thereby lowering the risk of recurrent subluxation or dislocation.²² Consequently, this surgical realignment improves patellofemoral tracking mechanics and decreases the compressive forces on the lateral patellofemoral cartilage by redistributing contact pressures medially.^23,24 This raised concerns regarding potential postoperative chondral lesions on the medial facet; however, our findings demonstrated no significant association between their development and medialization. Indirectly, there might also be an influence on the tibial-femoral component due to the altered knee biomechanics. Medialization of the tibial tubercle shifts the insertion point of the patellar tendon medially, resulting in an increased medial force vector, potentially increasing varus stress across the knee joint.^22,25 Ultimately, this leads to elevated contact pressures on this side, possibly contributing to its degeneration,²⁶ as observed in our study. We were unable to identify any specific risk factors associated with the development of chondropathy in our cohort.

TTO seems to be, in fact, effective in preventing further dislocations or subluxations, but its role in PF degenerative changes is still yet to be clarified.²⁷ The phenomenon of articular cartilage wear and chondral injury is poorly understood, and the available literature is not categorical concerning the predictive factors for PF osteoarthrosis and whether surgical management can speed up the process.^28,29 Although no significance was found, as potential risk factors for PF joint increased deterioration, we can point out older age and abnormal maltracking and alignment, translated by a high TT-GT distance, patellar tilt, and displacement. Delayed treatment may lead to chronic instability, characterized by subluxation or dislocation episodes, that can predispose to bone marrow edema and progressive chondral lesions. Early surgical intervention before the existence of significant osteochondral damage is therefore recommended to mitigate long-term joint degeneration.^11,30–32

Our study has a recurrent patellar dislocation rate of around 14 %, consistent with the literature-reported ranges of 5–20 %.³³ All episodes occurred in young female patients, corroborating epidemiological data identifying female sex, adolescence, and high physical activity as key risk factors for instability recurrence.^19,32 Although the rate of re-intervention for screw removal was relatively high, it did not appear to affect final functional outcomes negatively.

Our long-term outcomes show that TTO is effective in preventing recurrent dislocations, with both PROMs indicating high patient satisfaction and a low complication rate, corroborating previous reports in the literature.^6,19,30,34 Most importantly, no clinical decline was identified over time, different from other papers by Nakagawa³¹ and Carney,³⁵ even with the development of chondral lesions. As far as we know, this is the first study that, in addition to combining clinical and radiological evaluation, also aims to understand the role of PF osteoarthritis development.

Our study is not without limitations. The first limitation stems from the retrospective nature of our paper, which is inherently prone to certain biases. We addressed this by ensuring a long follow-up period and carefully selecting a uniform patient cohort based on strict inclusion and exclusion criteria. The second one involves a potential bias in comparing preoperative and postoperative patellofemoral measurements. Preoperative assessments were conducted using CT scans, as this was the standard practice at the time, whereas postoperative evaluations used MRI to reduce radiation exposure. While these imaging modalities differ, both are validated for precision and accuracy in this context. To minimize this discrepancy, we adhered to consistent evaluation criteria across all patients regardless of the imaging technique. Thirdly, we acknowledge that the absence of data on the patellofemoral joint condition at the time of surgery can be a limiting factor. This makes it unclear whether our findings reflect the progression of osteoarthritis or the impact of pre-existing chondral lesions. Despite a thorough review of available imaging, such as X-rays and CT scans, these data gaps remain a challenge. To build on our findings, future research should prioritize long-term, prospective cohort studies to better evaluate outcomes and refine treatment strategies in this field.

5 Conclusion

Tibial tubercle osteotomy offers consistent long-term outcomes for patellofemoral instability, however, the risk of joint degeneration, particularly patellofemoral chondropathy, remains a concern, especially in patients with significant malalignment or when the surgery is performed late.

Conflict

The authors have no disclosures to make nor conflict of interests.

No funds were involved in the process of this paper.

Declaration

No funds, grants, or other support were received.

The authors have no relevant financial or non-financial interests to disclose.

Credit author statement

Ana Rita Senra and Maria Clara were independent reviewers. They assisted in developing the methodology, extracting and analysing data, and working to diminish publication bias.

Paulo Oliveira and Francisco Serdoura were listed as co-workers because their expertise is knee pathology; likewise João Torres, who aid with the writing and correction of the discussion. They also helped formulate the research question and define inclusion and exclusion criteria.

Maria João Leite was responsible for ensuring that a rigorous methodology was followed and correcting all grammatical structure, helping in scientific writing and revising the manuscript.

This being said, all authors reviewed the manuscript, and gave valid input.

Sample CRediT author statement

Ana Rita Senra e Maria Clara Correia: methodology; validation; investigtion

Maria João Almeida: writing – original draft; resources; validation.

Paulo Oliveira: writing – review & editing.

Francisco Serdoura: writing – review & editing.

João Torres: visualization; supervision; validation; data curation.

No funds were involved in the process of this paper.

Consent

No guardian consent was needed since all patients enrolled were older than 18 years.

To get the approval of the ethical committee, patients need to consent to their enrollment in the study.

Funding declaration

No funds were involved in the process of this paper.