To stretch - not to stretch, that is the question.
Does Cochrane solve one of the biggest physio-dilemmas?

Hi, this is Joanna Tokarska and Tomasz Jurys, that is one of the more unobvious duos of the Polish physiotherapy scene. We decided to join forces and tackle one of the most important texts of recent years, the Cochrane study on the effectiveness of stretching. It wasn’t easy, there were sweat and tears, there were borderline moments and boundaries we never thought of crossing before. Thus, we put into the hands of Polish physiotherapists a study that will still be talked about… 

What's the deal?

In 2017, an update of a systematic review of the literature on two common issues in physiotherapy was published, namely (verbiage here): stretching and contracture!!! Tadaaa!!! The aforementioned terms have broad meanings, and their use is fraught with controversy, much like the choice between Pepsi and CocaCola.

Authors Harvey et al. asked a question that still has no clear answer: whether to stretch these contractures or not to stretch these contractures. The study is extensive and applies to patients with various pathologies; we will focus exclusively on neurological patients.

However, before we move on to try to answer the question stretch or not? let’s pause for a moment to clarify some things. First of all, let’s establish what the authors meant when they wrote contracture.

Unfortunately, here we encounter the first major doubt: although Harvey et al. give an abbreviated definition of contracture in the introduction “The term is used to refer to structural changes in the muscle‐tendon unit and other soft tissue structures overlying joints which together limit joint mobility”,  lumping into the so-called “one bag” – neurological conditions – is not a good idea.

One bag – neurological conditions – of patients with stroke, spinal cord injury, cerebral palsy, or even Charcot-Marie-Tooth disease, seems not very legitimate. It is possible that the clinical picture of contracture is similar in all these patients, but does the course of different diseases, their dynamics and (above all) their different substrates allow us to generalize conclusions, recommendations, guidelines? Well, probably not really…

In neurological diseases, contractures are a real challenge – with this we all agree – from the difficulty of daily care (hygiene, dressing) to the lack of effect of physiotherapy: contractures can be so problematic.

What is a contracture and what can it be confused with?

Theoretically, every physiotherapist in the course of his work has encountered a patient who had a contracture. This problem occurs in the course of a very large number of conditions – both neurological and other. A contracture is defined as a restriction of the range of motion.

However, is it really that simple? Rather not, because when we recognize a contracture, we also expect the presence of pathological structural changes, that is, changes within the tissues that make up or surround the movement segment in question.

These changes can range from abnormal length of muscle fibers, to changes in the structure and length of tendons, to changes in connective tissue, fascia, or skin (such as in burns) – the sky is the limit. However, the common denominator of the listed clinical situations is structural change.

This is where things start to get interesting – every structural change should have a cause. Finding this cause would make it possible to influence it and the problem of contractures would be solved. However, this is not the case, as anyone who has worked even one day with patients with neurological conditions in the chronic phase of the disease knows.

Let it be a consolation to us physiotherapists that scientists don’t know what the exact pathophysiology of contractures is either. Seriously. Of course, the general causes of these structural changes are known, such as tissue inflammation (e.g., in rheumatic diseases), scarring (e.g., after surgery), or long-term immobilization, but the detailed pathophysiology of these processes still remains unknown.

In neurological conditions, the influence of spasticity and dysregulation of tension are indicated as the cause of contractures.

Spasticity reduces the possibility of movement, and this restriction leads to changes in the structure of the tissues (because they are not in the position of full range of motion), which consequently leads to the development of contractures. 

However, one very important question should be asked here: is every spastic body part affected by contracture? Well, no! It is possible that the range of motion is preserved and, despite the observed spasticity, the mobility test gives a normal result. Why is this important? 

Because we have a different effect of movement treatment in the situation of healthy tissues and defective signals from the Central Nervous System, and a different one – when the tissues have already rusted (that is, structural changes are present).

The situation we are discussing in the literature review is precisely the treatment of rusty tissues. That is, to put it simply: it is an attempt to answer the question of whether, through stretching, it is possible to stretch tissue that is already pathologically altered in people with neurological conditions (but not always with spasticity – just such a flip throughout the story). 

Stretching is unequal to stretching but is still stretching

Harvey et al. defined stretching as “techniques involve the mechanical elongation of soft tissues for varying periods of time”

You stretch a patient’s foot tissues for 3 minutes with your own hands – it counts! 

You put the same foot in a cast and change it in a week – it counts! 

You put a pillow under the foot overnight – you won’t guess – it counts!

Figure 1: Results of a survey published on the Facebook group Fizjopozytywni. 181 participants answered the question: what do you understand by the word stretching?

Creating a literature review is a bit like cooking oatmeal. When you think of oatmeal, you imagine a uniform mush consisting of cereal and milk. In this case, it also contains nuts, raisins, flax seeds and even a roughly sliced banana. In a word – it is not homogeneous. Statisticians would say that its homogeneity is unobtrusive, but about that later.

What does our oatmeal consist of and what is the nutritional value of its ingredients?

The described interventions can be divided into two groups. The first includes those whose use has NOT shown a statistically significant effect in patients. That is, these procedures do not work, or at least their beneficial effect was not found in the papers included in the review. 

In the second group we will include an intervention (let’s not kid ourselves – there will be only one here) whose effectiveness can be considered statistically significant, although clinically it is not quite so….

A sad list of interventions not supported by scientific evidence:

  • (kinesio) taping  
  • orthoses/splints
  • positioning
  • other stretching methods maintained over time.

Yes, we know that your clinical experience may contradict the facts cited. We also realize that you are probably using some procedures not included in this publication, such as combining several methods at once, or conducting therapy in ways not mentioned in this review at all. 

This is why it is so important to document your activities and share your observations with others! If we don’t conduct research and publish it, we will forever remain shamans on a flying carpet. Reliably conducted and published research helps other physiotherapists make clinical decisions.

 Seriously, it’s worth spending some time doing research!

Now let’s return to stretching and its ins and outs: indeed, one type of intervention showed a small – but still – statistical effect.

That intervention was self-therapy, or self-stretching. But let’s not get excited: firstly, the average improvement oscillated around a few degrees (specifically, three degrees), and secondly, it involved patients who did not have neurological conditions.

While humming Tell me why… Backstreet Boys under our breath, we decided to become intrepid travelers on the ocean of obscurity and asked the question to the author herself. As soon as we get an answer, we’ll post it in this post.

Why it's important

Despite its flaws, the published literature review is extremely important. It is one of the few scientific texts giving strong arguments that allow for a serious discussion (academic, but especially clinical) about the validity of stretching as it has been practiced so far. 

And in particular about the validity and safety of stretching in neurology. Unfortunately, the literature review discussed here does not allow us to draw as clear-cut conclusions as we would like.

 Nevertheless, we will guide you through this 170-page text, the meanderings of statistics and compiled results, as well as the pitfalls set for us by the authors of the included studies, who are strenuously trying to prove their beliefs. 

We will do all this for you, so that you can draw your own conclusions on the basis of which you will make clinical decisions.

What are the study's findings?

Harvey et al. dug through almost all significant databases of scientific articles. They subjected each publication to strict inclusion and exclusion criteria, resulting in a final analysis of 28 studies with a total of nearly 900 participants. These 28 studies focus specifically on neurological diseases. What was the review authors’ goal?

 They evaluated, among other things, the effects of stretching on: range of motion (21 studies), pain levels (9 studies), quality of life (as many as 0 studies on the effect of stretching on quality of life), activity limitations (13 studies) and limitations in social participation (1 study, but no data to interpret).

 We clarify – we know that 21+9+0+13+1 does not equal 28, but a single study may have included several of the aspects mentioned (you know this too, but we want to avoid any doubt). Unfortunately, not all of these 21+9+0+13+1 studies could be included in the numerical analysis, as some data did not lend themselves to such interpretation.

Okay, okay, but to the shore Mr. Captain!

We know that at this point in your reading you would already like to know what the unequivocal conclusions of the literature review are, but the situation requires us to make a few more clarifications. 


We have already told you earlier about certain methodological aspects of this publication: the variation of disease entities, the incompleteness of the data, and so on and so forth… Therefore, these conclusions should also be approached with caution, but about that later. Now – tadaa!!!

Here you go:

Tab. 1. Quick conclusions for the lazy who are only interested in numbers.

Please admit that a 2° improvement in joint range of motion (with a five-degree measurement error on the goniometer – just a reminder) is not a crazy good score.

Other aspects we already mentioned were analyzed: activity limitation and reduced social participation; results were expressed in SMD (standardized mean difference). 

No, don’t be afraid, we won’t give you a mega lecture on statistics now! We’ll just point out that if the SMD value has a range of 0-0.2, there is virtually no therapeutic effect. 

And most of the results presented are precisely within such limits. 

What can go wrong?

Is stretching safe? Is it possible for something to go wrong during this therapy? Yes! As we have already pointed out, the study summarizes the results of as many as 28 studies on people with neurological conditions. Only 5 of them include any information on possible side effects. 

Damage to skin continuity, bruising, imprints were reported. Several children with cerebral palsy who underwent taping were withdrawn from the study because of more frequent falls. Could the fact that only 18% of studies included side effects of stretching interventions indicate that there were actually more side effects of the therapy?

 There is such a possibility. Here we express our respect to the authors of the literature review for being vigilant and paying attention to this aspect of treatment.

Is this literature review OK?

Results by results, statistics by statistics, tables by tables, but based on what we have learned so far about this study, can we consider it credible? Reliable, that is, one whose results can form the basis for formulating a specific clinical decision.

Before we try to answer this question, we would like to add a handful of our observations about the review in question. At the outset, we drew your attention to the fact that the authors included patients with multiple neurological conditions in one group. 

We consider such a procedure to be not very correct, since therapy through stretching may look different in individual disease entities. The correctness of our claim must be supported by certain facts. For example, in the case of acquired brain injury, the presented combined results of 3 studies evaluating short-term treatment effects show that stretching improves the range of motion in the joint by almost 8.5° – a statistically significant value (and we are tempted to say that it is also slightly clinically significant).

 All fine, but… This but is called the level of heterogeneity (I2) – we promise: this will be the last statistical term cited. Basically, statisticians say that if the level of heterogeneity is above 50%, the combined analysis of the results is… not very reliable; in fact, no such analysis should be done at all. Do you want to know what the I2 value was for the aforementioned 8.5° result? 63%.

 Well, that’s just it… Unfortunately, the literature review in question provides a lot of similar doubts. Such a situation occurs, for example, with the distribution of stretching interventions. That is, let’s summarize – if there is a result with a statistically significant value, even a clinically relevant one, we still don’t know what it actually means and whether it is reliable

You would probably like to ask yourself and us whether, based on what we have said so far, we can consider this literature review valuable?  

We answer emphatically: YES!!!


  1. The authors report everything, including statistically or clinically insignificant results – this means a high level of research objectivity;
  2. The scientific data are reliably collected – especially the effect of stretching on range of motion;
  3. The review appeared in the Cochrane Database of Systematic Review, and that speaks for itself. Only the best studies are placed there.

What are the implications of this for us physiotherapists?

The summary of the literature review in question reads: “stretch, as typically provided by physiotherapists, does not produce clinically meaningful effects on severity of contractures in people with neurological conditions”.

 That is, to put it simply, it doesn’t work.

The authors also emphasize that stretching is unlikely to reduce pain because the therapy does not affect range of motion.

Does this mean we should remove all forms of stretching from clinical practice? No positioning, no stretching, no teaching of self-therapy?

No! And this is not just the position of the authors of this text.

It should be remembered that the lack of improvement in mobility does not necessarily mean the failure of therapy, and in some cases, despite the apparent ineffectiveness of therapy, one can even talk about the success of treatment.

In many neurological conditions there is a gradual deterioration of patients’ functional status due to, among other things, increasing contractures, so the absence of such deterioration is definitely a success.

The authors themselves write in their conclusion it may be reasonable to use stretching regularly in people with chronic neurological conditions, over the course of their lives, to treat and prevent contractures; however, it is not known whether this is effective. 

What are your experiences? Stretching – yes or no?

Authors: Joanna Tokarska, Tomasz Jurys

error: Alert: Nie kradnij treści, stwórz swoją!!