Song et al. (2023)

Patellofemoral Joint Loading Progressions - Picking the Right Exercise at the Right Time

This study examined patellofemoral joint loads across 35 rehabilitation and everyday activities and 3 tiers of exercises were defined: low, moderate, and high loading

The patellofemoral joint loading progressions can be used to individually adapt the loads for patients with patellofemoral pain and anterior knee pain

Many fast, high-intensity exercises, such as running and single-leg fast forward hops, produce just a low or moderate amount of total loading, despite being regarded as high-intensity rehabilitation exercises


Last week, our research review covered the topic of running adaptations that can help manage patellofemoral pain. This article specifically focussed on runners. Despite patellofemoral pain being active in younger and often active individuals, not all are runners. This week’s research review tackles the topic of patellofemoral joint loading progressions to progressively put more or less demands on the patellofemoral joint. Knowing where to start and how to progress is especially important for you as a clinician to effectively guide patients with patellofemoral pain through the rehabilitation process. The reason these authors conducted this research was because recent rehabilitation programs may be too simplistic according to a systematic review by Dischiavi et al. (2021). The balance between too-low and too-high loads must be considered appropriately during rehabilitation to know where to start and when to reduce loads. How can you do this effectively? Read on!



For researching patellofemoral joint loading progressions, the study recruited healthy adults between 18 and 40 years who had no evidence of lower limb injury or spine problems for the last 6 months. They did not experience anterior knee pain, nor did they have a history of patellofemoral joint injuries.

These participants were given standard exercise clothing and running shoes and were set up with 31 markers over anatomical landmarks on the pelvis, upper torso, and legs. They completed 35 weight-bearing exercises commonly used in daily life or physiotherapy rehabilitation.

Then knee joint flexion angles and extension moments were calculated to define the loads on the patellofemoral joint. The estimated patellofemoral joint forces were normalized to the participant’s body weight. Three different loading variables were measured to quantify the loads on the patellofemoral joint:

  • Loading peak: which is the maximum force during each repetition
  • Loading impulse: the relationship between force and time
  • Loading rate: the maximum change of force over time
Patellofemoral joint loading progressions
From: Song et al., Am J Sports Med. (2023)


The mean of these three variables was used to calculate the loading index, which varies between 0 and 1 with 0 representing no load and 1 representing an exercise that would have a maximal loading peak and impulse.

The exercises were ranked into 3 tiers to allow making patellofemoral joint loading progressions.

  1. Tier 1 were exercises with a loading index below 0.333: low loads
  2. Tier 2 reflected a loading index between 0.334 and 0.667: moderate loads
  3. Tier 3 consisted of exercises with a loading index exceeding 0.667: high loads



The figure hereunder displays the ranking of the exercises according to the loading indices. The lowest loads on the patellofemoral joint were exerted during walking. The exercises in tier 1 had a peak patellofemoral joint loading ranging from 0.6 to 4.9 times body weight.

Tier 2 comprised the most exercises, the lowest loading exercise was a bilateral repeated lateral jump at a regular speed. This exercise was just below the loading of running. The highest patellofemoral loading exercise from tier 2 was the unilateral countermovement hop. The peak patellofemoral joint loading in this tier ranged from 4.3 to 7.1 times body weight

Three challenging exercises were in tier 3:

  1. Single-leg squat at full depth
  2. 3-second Spanish squat
  3. Single-leg decline squat

The tier 3 exercises ranged from 4.5 to 8.2 times body weight.

Patellofemoral joint loading progressions 2
From: Song et al., Am J Sports Med. (2023)


This figure is constructed based upon a 50% weight of loading peak and a 50% weight of loading impulse. The authors provided an Excel sheet in the supplementary material that you can adapt to re-rank the exercises and to alter the weight of the loading peak, impulse and rate.

The table below shows the peak loading, impulse and rate loading for all exercises across all three tiers.

Patellofemoral joint loading progressions 3
From: Song et al., Am J Sports Med. (2023)


Questions and thoughts

This study is of particular interest since it only studied weight-bearing exercises to make patellofemoral joint loading progressions. Since exercise loading for patellofemoral pain is frequently insufficient, this study opens up an important door.

Loading peaks, impulses and rates varied much: a jump produces a larger peak than a double leg squat for instance, although the latter is characterized by a larger loading impulse because of the longer duration. Similarly, the run-and-cut exercise and the Bulgarian squat had a similar loading index. With the loading index being equally high between those two exercises, the difference again lies in the duration of the exercise and the peak loads. Therefore it is important to think about the stimulus you’d want to train during an exercise. Here the Excel worksheet provided by the authors is especially useful as you can re-rank the exercises according to what stimulus you want to accentuate.

For instance, a physiotherapist may view loading rate as a main parameter given the mechanical properties of articular cartilage (which is loading-dependent) and the fact that patients with patellofemoral pain frequently find it difficult to resume running. As a result, the clinician might arrange these workouts in a different sequence that gives priority to a gradual increase in loading rate. This can be done by giving the exercise more weight in the loading index.

Patellofemoral joint loading progressions 4
From: Song et al., Am J Sports Med. (2023)


The three Level 3 exercises are all squat variations and require a great deal of knee flexion. They also take longer, adding to the strain on the patellofemoral joint. Squat variations are often performed before jumping exercises. Contrary to the commonly held belief that jumping and running put heavy stress on the knee (and the patellofemoral joint), this study shows that because of the longer duration of the forces exerted during squatting, the stress on the joint is much greater than during jumping and running exercises.

When the figure was rearranged to put more accent on the loading rate (50%) of the exercise (the maximum instantaneous change of force over time), less peak (30%) and less loading impulse (20%), the order of the exercise changed. For example, a run-and-cut exercise that was classified as a tier 2 exercise became a tier 3 exercise. The reason lies in the rearrangement of the patellofemoral joint loading indices. A run-and-cut is a fast and dynamic exercise and requires less loading over time (loading impulse) than a 3-second Spanish squat. When less weight is given to the loading peak and impulse and more to the loading rate, a longer duration exercise will become an exercise creating more loading rate and as such will be re-ranked to a tier 3 exercise.

Patellofemoral joint loading progressions 5
From: Song et al., Am J Sports Med. (2023)


Talk nerdy to me

The exercises were divided into 4 modules to minimize physical exhaustion and suboptimal exercise execution. You can see the description of each exercise in Appendix A1 on the publisher’s site. Seven to ten trials were recorded with various levels of intensity with 3-5 minutes rest in between each exercise. It was visually verified that the participants did not get exhausted. Furthermore, the loads were normalized to body weight to allow the comparison between exercises and people of different weight categories. These are all good methodological aspects.

Some limitations may lie in the following:

  • The exercises were executed in the same order. This may have caused a learning effect.
  • A model was constructed that restricted the knee to only flexion and extension, which is of course too simplistic. Therefore, this study cannot make assumptions about what is happening in the frontal (abduction and adduction) and transverse planes (internal and external rotation) of movement.
  • The right-sided leg was chosen in the analyses, while no was given to the leg preference. I can imagine that if your left side is dominant, you’d perform better when your left leg is measured.
  • The knee flexors were not considered in the model, and as such, the potential effects of knee co-activation are neglected. Quadriceps force may be underestimated up to 1.5 times bodyweight when neglecting muscle coactivation.

Powers et al. (2014) reported the peak patellofemoral joint loading stress of an open chain leg extension to be equal to the 60° double-leg squat which was classified as a tier 1 exercise in this study. The authors of this study emphasized that when you perform an open kinetic leg extension exercise with more knee flexion, the peak patellofemoral joint loading is even less than the loading experienced in the joint during walking. This again confirms the safety of the often-shamed open kinetic chain knee extension exercise. Be sure to check out Kai’s blog in case you want to learn more about that!

An important aspect to mention is the healthy population in which this study was carried out. These patellofemoral joint loads may therefore not be directly representative of someone with patellofemoral pain. However, by studying healthy participants without knee pain, the influence that pain might have on the performance of exercises is removed and we can assume that the ranking of exercises may be representative of what happens in the knee joint independent of pain.


Take home messages

This study provides us with 35 exercises ranked for their relative contribution to patellofemoral joint loading progressions. If your patient experiences difficulties with running for example, you can use this information to adapt the rehab by temporarily picking exercises that create less patellofemoral joint loading. The exercises can be ranked according to the relative weight you want to assign to someone’s rehabilitation. For some, you’d want to create more loading over time instead of a loading peak (for example activation during early phases of rehab), for others you’d prefer creating more loading peaks during the exercises (for example late stage rehab and return to sport phases). The online Excel worksheet lets you rearrange the exercises according to your rehabilitation preferences. As such, this study can be used to make patellofemoral joint loading progressions based on evidence, rather than on your best guess!



Song K, Scattone Silva R, Hullfish TJ, Silbernagel KG, Baxter JR. Patellofemoral Joint Loading Progression Across 35 Weightbearing Rehabilitation Exercises and Activities of Daily Living. Am J Sports Med. 2023 Jul;51(8):2110-2119. doi: 10.1177/03635465231175160. Epub 2023 Jun 5. PMID: 37272685; PMCID: PMC10315869. 



Watch this FREE 2-PART VIDEO LECTURE by knee pain expert Claire Robertson who dissects the literature on the topic and how it impacts clinical practice.


Vmo lecture
Download our FREE app