2005 Review. The PEDro scores among the 28 studies included in the original review ranged from 2 to 8, with a mean of 4.3 (SD = 1.6). Several factors, both positive and negative, stood out during that review: 93% of the studies reported point measurements and variability (usually means and standard deviations) associated with their results; 86% reported the results of their statistical analyses; 82% assigned participants to intervention groups randomly; only 18% of outcome assessors were blinded to group assignment. Two other criteria, which are significantly related to each other, were also areas of concern. Criteria 8 and 9 regarding the minimum 85% follow-up and intent to treat analysis were satisfied in 57% and 36% of the studies, respectively. It is important to note that satisfaction of these criteria requires explicit reporting related to these items and that reporting was absent; of course, this could mean that there was inadequate follow-up or it could mean that it simply was omitted from the article. One17 of these 28 studies met or exceeded the 70% methodological score suggested18 for recognition as a valid clinical trial.
2009 Review. PEDro scores among the 11 studies reviewed for this publication range from 4 to 7, with a mean of 5.7 (SD = 1.1). This reflects improvement in the methodological quality of studies published more recently. Criteria that were satisfied more frequently in the current review include random assignment of participants to groups improved from 82% to 100%; evidence that the groups were similar at baseline improved from 57% to 91%; assessor blinding improved from 18% to 36%; follow-up and intent to treat analysis criteria improved from 57% to 82% from 36% to 82%, respectively. Two11,15 of these studies met the 70% methodological score mentioned above; clearly 2 of 11 is better than 1 of 28; however, overall quality is still lacking.
Impact of Stretching Position, Technique, and Duration on Flexibility Gains
2005 Review. The clearest finding among the studies was that nonstretching control groups do not gain range of motion. Eleven studies17,19–28 reported point measurements for their control groups; gains ranged from −3° to 2.7°, and no study reported a significant change from baseline for the control group. A similar general statement can be made referring to the stretching groups—they do gain range of motion. It was impossible to identify the best stretching protocols for most parameters. Highlights from the previous review follow.
Among the 28 studies, hamstring flexibility gains ranged from 5.7° to 33.6°. Whether the hamstring stretch was performed standing, seated, or supine did not appear to make a significant difference in the magnitude of flexibility gained. None of the studies reviewed was specifically designed to compare gains based on stretch position.
However, there were studies in the previous review that directly compared stretching techniques. Two studies29,30 were designed to compare static and proprioceptive neuromuscular facilitation (PNF) techniques. In the first,29 static stretchers gained 9° and PNF stretchers gained 12°, a statistically significant finding. However, in the second study30 there was no difference—static and PNF stretchers gained 8° and 9.5°, respectively. Bandy et al21 found static stretching to be significantly more effective (11.4°) than a dynamic knee extension exercise (4.3°). Wiemann and Hahn28 found no significant difference between static (7.8°) and ballistic (8.4°) stretching in a single stretch session.
Stretch duration was a topic of direct comparison in four19,20,31,32 of the previously reviewed studies. Bandy et al19,20 have been responsible for 2 studies that have concluded that 30 seconds is an ideal length for stretching, with gains ranging from 10.1° to 12.5° regardless of whether 1 repetition or 3 repetitions are performed. Comparing nine 5-second stretches to three 15-second stretches, Roberts and Wilson32 found that the 15-second stretch gained more range (7.8°) than the 5-second stretch (4.6°). In another study comparing the effects of overall stretching times, Cipriani et al31 compared six 10-second (28°) stretches with two 30-second (24.2°) stretches and found no difference.
Another duration variable of interest is the length of the stretching protocol (a single session versus multiple sessions over weeks). Unfortunately, although there was a wide variety of protocols ranging from single sessions17,22,26,28,33–37 to twice per day for 6 weeks27 and even as long as 10 weeks,29 only the method used by Cipriani et al31 included this comparison; they found that significant gains had been made by 3 weeks (gains continued in the ensuing 3 weeks). A superior protocol could not be gleaned from the 28 studies; all showed gains. The length of benefit (ie, how long the increase in flexibility lasted) was also not commonly studied. DeWeijer et al17 showed continuing significant improvement 24 hours after a single session of stretching; others showed significant benefits 1 day38,39 or 2 days19–21,24 after the end of the stretching protocol.
2009 Review. From the current review, it can again be concluded that nonstretching control groups did not gain range of motion. Five10,12–15 of the 11 studies included control groups with range of motion gains ranging from −3.2° to 3.2°. All but two6,7 of the studies reported range of motion gain as an outcome variable with results ranging from 1.9°13 to 23.7°.10 Although most flexibility increases were within the range reported in the previous review, 2 studies had very low, albeit reportedly statistically significant, gains. Cronin et al9 and Ford et al13 reported gains between 1.9° and 3.6°, more comparable to gains associated with control groups. Certainly, the clinical significance of gains so minor could be questioned.
Of note, in both studies with relatively small gains, the stretching position used had not been examined in any of the other 37 studies reviewed. Cronin9 had participants in a lunge-like position with the back of their thigh resting on a padded surface; the stretch involved active knee extension from that start position. Ford13 had participants sitting in a chair with the leg to be stretched extended in front of them, resting on its heel; the stretch involved assuming an anterior pelvic tilt and leaning forward over the leg. Again, in the current review, no strong conclusions can be made about the effects of stretching position. Among the more recently published studies, 1 study11 was designed to compare stretching positions. It concluded that there was no difference in range gained whether the static stretch was performed in a standing (9.4°) or supine (8.1°) position.
Three studies7,10,14 directly compared stretching techniques. Comparisons included static versus ballistic stretching7 and active, static, and PNF.10,14 No significant differences were identified in resulting flexibility, although significant differences between control groups and stretching groups were found in the 2 studies that included control groups.10,14
Two studies 8,13 were designed to compare stretch duration. One8 compared 3-, 6-, and 10-second holds during hold-relax PNF stretching. The other13 evaluated 30-, 60-, 90-, and 120-second holds during static stretching. Neither study identified significant differences in flexibility gained regardless of the stretch hold duration; Ford et al,13 like Bandy et al,19,20 concluded that 30 seconds is an adequate stretch time.
None of the studies was designed to compare protocol duration and length of benefit measurements. Only one collected post-stretch data even as far out as 25 minutes,14 so there is not much to add to the previous review’s consideration of the differences that those 2 parameters might create. Three of the studies were designed to assess the effects of associated warm-up6,16 or modalities.9 Much like the 2 studies17,28 in the previous review that included a group that performed the warm-up but no stretching, the 2 studies in the current review concluded that stretching created significantly more range of motion gain than did warm-up alone, and that adding the warm-up to the stretching protocol did not improve gains. The design of one of those studies6 allowed the conclusion that placement of stretching before or after activity did not affect range of motion gains. Cronin et al9 found there was no flexibility benefit associated with vibration therapy.