Effectiveness of Proprioceptive Neuromuscular Facilitation (PNF) – Evaluation of Literature

To evaluate the evidence of the effectiveness of PNF in post-stroke adults, I conducted a search of literature in online databases such as Ovid, Sciencedirect, Cochrane Library, CINAHL and Wileyinterscience. The search included keywords such as “proprioceptive neuromuscular facilitation”, “physical therapy” and “stroke”, with Boolean operators being used to combine the keywords thus restrict the results to the most relevant ones. Further searches were conducted after manually scanning references that were predominant in the preliminary studies identified for review. Although, initially, the search was limited to articles published from the year 2000, such limitation was dropped following the scarcity of studies conducting primary research within this range. Additionally, recent reviews (e.g. in Pollock et al. 2007 and Westwater-wood 2010), feature core studies that were not within this range thus the limitation of publication date was dropped to gather evidence from these primary sources. This paper thus reviews the evidence as provided in studies carrying out primary evaluations and recent systematic reviews on the topic.

One of the studies that highlighted the efficacy of PNF was a comparative study of three approaches (conventional treatment, PNF and Bobat approach) conducted by Dickstein et al. in 1986. Using a sample of 131 patients, drawn from an initial sample of 196 hemiplegic patients, the study compared the efficacy of the three approaches using various measures. Firstly, using the Barthel index (a measure of activities of daily living; Ashburn 1997), the study assessed any functional gains obtained in fostering activities of daily living following each treatment. Secondly, via a five-point ordinal scale, the researchers evaluated the effect of the treatments on the muscle tone. Other outcome measures used were effects of the treatments on isolated motor control and such treatments’ effects on the patients’ ambulatory status. Although no significant differences were observed in most of the outcomes among the three approaches following a six-weeks therapy, PNF did enhance, though not significantly, the pattern of the muscle tone to a greater extent as compared to the other approaches.

The strength of the study lies in its use of well-informed variables. For instance, in a pilot study, the authors had conducted found the interrater reliability for all criterion variables to be above 0.9. Independent studies have also confirmed the validity and reliability of the Barthel study as a measure of disability (Wade & Collin 1988). Additionally, the study also describes the activities conducted under each of the treatment approaches, a factor that lacks in other studies especially with respect to the “traditional/conventional” approach (Ernst 1990). However, the study’s weakness arises out of limitations imposed by the use of ordinal and nominal scales. Although common within clinical settings, such scales are prone to introduction of subjectivity in the assessments and they lack clear discriminative power. For the current study, the risk could be aggravated since evaluations were non-blinded.

Ernst (1990) in a review of the physiotherapeutic approaches in stroke rehabilitation also reinforces the failure to identify an optimal approach. Despite such failure, the review notes that, albeit lacking a clear link in literature, physiotherapeutic approaches do benefit stroke patients. Accordingly, the review poses that, until further evidence to delineate an optimal approach is presented, selection of therapies is enhanced by considering how cost-effective an approach is and the largest number of patients that would receive treatment following adoption of the respective approach. Although the study provides insight on the need to consider the cost-effectiveness of an approach, the findings are subject to the studies reviewed and limited to author’s evaluation of these studies, since the review is not a meta-analysis.

A study by Wang (1994) provides support for the use of PNF. Two objectives guided the study: (a) “to compare the immediate and cumulative effects of PNF treatments applied to the pelvic region in patients with hemiplegia” and (b) to evaluate whether the duration (short or long) of hemiplegia altered the treatment outcome (Wang et al. 1994, p.1109). Short-duration hemiplegia ranged from 2.8-5.6 months and the long duration one ranged from 12.7-18.5 months (Wang 1994, p.1109).

A pretest-posttest design was used with a convenience sample of 20 adults (12 men and 8 women). The inclusion criteria selected for aspects such as diagnosis of hemiplegia being secondary to cerebrovascular accident, an unassisted walking ability for at least 18.3 m and the patients were not to have received prior PNF treatments. Additionally, the patient, as evaluated by the therapists by tests on the hip, knee and joint regions, had to have intact kinesthesia for all the three regions (Wang 1994). The duration of hemiplegia and the demographics of the participants were obtained from interviews with the patients and medical records. The sample had 10 patients (6 men and 4 women) in either of the short- or long-duration hemiplegia categories. For outcome measures, gait speed and cadence were used to assess the change in gait following therapy. The therapy was conducted in 12 sessions (three times per week) for all the patients with each treatment followed by a walk of 18.3 m where the total number of steps and heel-strikes were counted. Participants were allowed two minutes of rest between each trial. Gait speed was calculated by dividing the distance walked by time, whereas cadence was obtained by dividing the number of steps by the time taken. Analyses included t-tests to evaluate differences in effect between the two groups and mixed-factorial analysis of variance to evaluate intra and inter-group variances.

The results indicated that patients benefited from PNF treatments. For the short-duration group, even a single session improved the gait speed and cadence with such improvement being enhanced by further treatment. For the long-duration group, however, a single session did not reveal improvements in the gait, but after the 12 treatments, an improvement similar to that noted in the short-duration group with similar number of treatment sessions was observed. Following such observations, it appeared that the duration of hemiplegia did not affect outcomes to sustained PNF treatment, but it did affect the speed of response to the treatment provided. However, lacking a control group, which did not receive PNF treatment, the observed effects of PNF treatment could have resulted from confounding factors.

Ferber et al (2002) also offer support to PNF’s efficacy in enhancing ROM in their study that evaluated the “effects of PNF stretch techniques and EMG activity in older adults” (p.392). The authors compared effects of static stretch (SS), contract-relax (CR) and agonist contract-Relax (ACR), in 24 elderly (50-75 years) normal subjects. The outcomes measures were knee extension ROM and EMG activity in the biceps femoris and gastrocnemius. The study findings were that ACR was more effective in enhancing ROM than SS and similarly its effect on EMG activity was higher compared to that of CR. Despite the finding that PNF increases ROM about a joint, the study observes that such effect was not correlated with enhanced muscular relaxation in the study sample. Thus, age could be a critical factor determining the outcome of PNF approaches. Use of EMG activity to determine effect of PNF, with respect to muscle length, has however been contested. Magnusson et al., for instance contend that changes in torque after treatment could be explained more by passive aspects than activity of neuron unit (as cited in Westwater-Wood 2010, p.25). This, they observe, results from the possible differences in tension exerted at baseline EMG-recording and post-treatment EMG-recording (as cited in Westwater-Wood 2010, p.25). Results obtained in the study may also not be replicated stroke-rehabilitation setting, since normal subjects used in the study could possess different profiles with respect to muscle strength and stretch as compared to post-stroke patients (Wang 1994).

A related study also compares effect of SS and CR on tone, elasticity and stiffness of the biceps femoris muscle. In this study, Rihvk et al. (2010) using a sample of 30 male subjects assigned to three groups – a control group, a SS group and a CR group – found out that CR had no effect on the tone, elasticity or stiffness of the muscle. The study’s strength lies in its randomized-controlled design that allows the comparison of effects of test groups with those of a control group. The study’s findings however may have been affected by the measurement approach employed to measure pre- and post-test values – the myoton-3 myotonometer – whose reliability has not been conclusively evaluated (Rihvk et al. 2010).

Shimura and Kasai (2002) also provide support for PNF’s efficacy by investigating how the approach “affects the initiation of voluntary movement and the momentary coupling of activities of several muscles” (p. 102). The experiment used a wrist extension task to compare the outcomes when the participants were assuming a PNF posture and when they were in a neutral posture. Initiation of the voluntary movement was obtained as a measure of the electromyographic reaction time (EMG-RT), whereas that for momentary coupling of muscle activity was evaluated via motor evoked potentials (MEPs). EMG-RT is a measure of the electrical activity of skeletal muscles (Gunendi et al. 2005), whereas MEPs are waveforms generated following stimulation of motor components of CNS (Chiappa 1997. p. 477). The results of the study indicated PNF to be a potential approach for use in rehabilitation of patients with impaired movement. For instance, when compared with neutral position, facilitation enhanced the joint’s movement efficiency by changing the muscle discharge order and reduced EMG-RT by a magnitude dependent on how close the muscle was to the joint being moved. Additionally, such facilitation increased MEP amplitude while decreasing its latency, in a pattern dependent on how close the muscle was to the joint being moved. The study’s findings may however be affected by the limitation of using EMG activity as an outcome measure in treatments where improving muscle length is one of the goals. Additionally, the small sample used may not provide adequate evidence to generalize the effect observed in other settings.

A recent study by Wolny et al. (2011) supports the use of PNF in rehabilitation following stroke. The study aimed to assess how spasticity differs between post-stroke patient treated with PNF approaches and those given traditional treatments such as psychological assistance and use of walking aids. The sample size was 64 stroke patients who were randomly assigned to the two treatment approaches, involving an intensive comprehensive rehabilitation program provided in a 21-days follow-up evaluation. In one of the groups, the kinesiotherapy provided was based on PNF approaches. A modified Ashworth scale, comprised of a 6-grade scale with 0 representing normal muscle tone and 4 coding for rigid flexion and extension in the affected limb segment, was used to assess the extent of spasticity. The modified scale has been found to be reliable in assessing muscle spasticity (Bohannon & Smith 1987). The study results indicated that both approaches had provided significant improvements of the patient’s condition, but patients who were given the PNF treatment had a greater improvement. For instance, whereas in the final examination, as compared to the baseline evaluation, the group provided with traditional treatment had a 1.8% improvement, such improvement in the PNF group was 16.8%. Even following the first session of PNF therapy, the gains the patients in the group made (7.8%), outweighed the cumulative gain observed in the group treated with traditional approaches. Accordingly, PNF could offer greater benefits in terms of reducing spasticity as compared to the traditional approach. The strength of the study rests in its randomized-controlled study design but it does not provide adequate statistics to evaluate whether the differences noted in effects following PNF and traditional treatment approaches are significant.

A different study provides evidence that the effect of PNF may be enhanced via its combination with electrical stimulation (Munih et al. 2004). In this study, the authors sought to compare the effect of such a combined therapy (of PNF and functional electrical stimulation; FES), with either of the approaches given independently. The study used two subjects who were monitored for one month. Four types of activities were provided including activities leading to voluntary movements, those generated by electrical stimulation following placing of two electrodes over the peroneal nerve and PNF-based activities (e.g. knee flexion and dorsiflexion). The final group of activities were designed to evaluate the effect of combined therapy and comprised of activities ranging from those aimed to generate voluntary movements by the subject, through those that led to electrically-stimulated movements, to a PNF-sequence applied by the physiotherapist. The results of the study indicated that the effect of voluntary exercises was almost equivalent to that of PNF activity in one subject but inferior to PNF-activity for the other subject. For this later subject, effects of voluntary activity were comparable to those achieved when only electronic stimulation was employed. When the combined therapy was considered, at most dates, it had a greater angular response for the hips than either PNF or FES activity. Such indications mean that a combined therapy may offer better efficacy, an aspect reinforced in other studies (Chen & Shaw 2006), but this study’s limitation of a small sample size prevents the generalizability of its results. By using two subjects, the differences observed in treatment outcomes may be due to individual characteristics rather than treatment effects.

Another study that evaluated the efficacy of PNF in a mixed therapy as opposed to stand-alone therapy compared traditional post-stroke treatments and PNF to a mixed therapy that, in addition to the two approaches, included Butler’s neuromobilizations. The study used 96 late-stage stroke patients who were randomly assigned to the traditional approach (control group), PNF and the mixed-methods groups, in a pretest-posttest study design. The measures evaluated the “distance between the compass tips when the subject indicated two-point sensation” (a two-point discriminatory sense), stereognosia and thermaesthesia (Wolny et al. 2010, p. 810). The last two aspects are a measure of touch sensation and that of perception of cold and hot sensations, respectively (Wolny et al. 2010). The study found out that in the mixed approach group, the two-point discriminatory sense and thermaesthesia were significantly improved as compared to the other two groups. With respect to stereognosia, the effects of the approaches were noted to be significantly different only between the control group and the mixed-approach group. Accordingly, such a combined approach could offer greater effectiveness in rehabilitation of post-stroke patients as compared to independent PNF treatment.

Two studies based on a review of previous studies also present equivocal evidence on the use of PNF. In one of the studies, based on a meta-analysis, Pollock et al (2007) aim to establish whether physiotherapeutic approaches based on neurophysiological, orthopedic, motor learning or a combination of the three concepts, offer different efficacy in respect to recovery of postural control and the function of lower limb in post-stroke patients. The study used studies from various online databases with the time period for publication of such studies ranging from 1966 to May 2005. Only randomized or quasi-randomized studies that examined adults with a diagnosis of stroke were included. The study did not find a significant difference to exist in the efficacy when neurophysiological or motor learning approach was used. Additionally, no single approach was found to significantly have a better efficacy than any other single therapeutic approach. However, the study found out that, when compared to a control group, a mixed approach to therapy offered a significant positive effect in regaining both postural control and lower limb function. Although the review provides a high level of evidence since it reviews only randomized and quasi-randomized studies, it only evaluates PNF as a component of neurophysiological approaches thus does not offer evidence that is specific to PNF.

A second review concentrated solely on PNF. In their review, Westwater-wood et al. (2008), sought to evaluate the evidence base for PNF’s use in physiotherapy. The review used studies generated from various databases. The studies included in the review examined various components that would elucidate the effectiveness of PNF including how PNF affected the ROM and functional rehabilitation. Other studies reviewed included those on implications of PNF patterns (e.g. Shimura et al. 2002) and one survey on physiotherapists’ perspectives on the efficacy of PNF. The survey, which was conducted in a large UK hospital, generally indicated the support for use of PNF in physiotherapeutic practice (cited in Shimura et al. 2002, p.26). Another independent survey conducted in Kansas and Missouri also provide evidence of support for PNF, with PNF coming second only to Bobath in its training at school and use in practice (Natarajan et al. 2008). From the review by Westwater-wood et al. (2008), the overall finding is that PNF effect is particularly found with regard to enhancing ROM, but due to the varying methodologies of the studies reviewed, the authors note of a limitation to compare the evidence from different studies. As such, they recommend the need for standardized measures that would enhance such comparison in specific contexts.

Supporting some of the arguments in the review by Shimura et al. (2008), Sheard (2010) also note of the ambiguity in some of the practices (e.g. those termed only as conventional or traditional approaches) against which PNF and other physiotherapeutic approaches are compared. Further, the wide-range of approaches that are considered within the term PNF (Sharman 2006), limits the extent to which its efficacy can be elucidated effectively. In this respect, Westwater-Wood et al. (2008) point out the need to evaluate individual components of PNF to delineate the approaches that are suitable for application in different contexts.

Go to conclusion here.

find the cost of your paper