As training progressed beyond the initial sessions, cognitive behaviors that influenced the pain behavior were also targeted (35). Participants who experienced pain-related fear of movement and were apprehensive about stressing their low back by lifting weights were individually reassured that, by maintaining the neutral position of the spine during daily activities, the risk of aggravating their symptoms would be minimized and the progression of strenuous activities made possible.
Means, standard deviations, and quartiles were used for descriptive statistics. To evaluate which individual factors could predict activity, disability, and pain intensity in patients with mechanical low back pain after 8 weeks of deadlift training, separate predictive models were constructed using linear regression with the follow-up values of activity, disability, and pain intensity as dependent variables. First, for each predictive model, univariate linear regression analysis was performed by testing each independent variable separately. The independent variables were the baseline values of activity, disability, pain intensity, pain-related fear of movement, the movement control test battery, the Biering-Sørensen test, the prone bridge test, the side bridge test, age, sex, and body mass index. Second, multiple linear regression analyses were performed that included the variables from the univariate analysis with p ≤ 0.05. The independent variable with the highest adjusted R2 from univariate analysis was first entered into the multiple regression models. Thereafter, the remaining significant independent variables from univariate analysis were added one at a time in order of relevance to gauge whether they could increase the explained variance (adjusted R2) and form a significant (p ≤ 0.05) regression model.
There was a high degree of adherence to training among participants (n = 35), 33 of whom completed the 8-week training period and attended 11.6 (SD = 0.8) of the 12 planned sessions. Two participants reported adverse effects, one of whom withdrew from the study. Another participant withdrew from the study without explanation. Both participants who withdrew reported baseline activity scores of 4.3 and 5, disability scores of 7 and 8, and pain intensity scores of 57 and 54 mm, respectively. For activity and disability, these scores were similar to the mean score of the group, whereas the score for pain intensity was considerably higher. Generally, activity scores increased from 4.8 (SD = 1.3) to 6.8 (SD = 2.2) (1), disability scores decreased from 7.1 (SD = 4.1) to 3.8 (SD = 3.9), and pain intensity scores decreased from 42.6 (SD = 23.5) to 22.2 (SD = 21.1) (1).
Prediction of Follow-up Scores for Activity (Patient-Specific Functional Scale)
The univariate regression models with the follow-up value of activity as the dependent variable showed associations with 4 independent variables: disability, pain intensity, the Biering-Sørensen test, and the prone bridge test (Table 2). The Biering-Sørensen test was entered first in the multiple regression model. No other variable except pain intensity could significantly increase the explained variance (adjusted R2 = 0.23) (Table 2).
Prediction of Follow-up Scores for Disability (Roland-Morris Disability Questionnaire)
The univariate regression models with the follow-up value of disability as the dependent variable showed associations with 3 independent variables: disability, pain intensity, and the Biering-Sørensen test (Table 2). Pain intensity was entered first into the multiple regression model. No other variable except the Biering-Sørensen test could significantly increase the explained variance (adjusted R2 = 0.17) (Table 2).
Prediction of Follow-up Scores for Pain Intensity (Visual Analog Scale)
The univariate regression models with the follow-up value of pain intensity as the dependent variable showed associations with 3 independent variables: disability, pain intensity, and the Biering-Sørensen test (Table 2). Disability was first entered into the multiple regression model. No other variable except the Biering-Sørensen test could significantly increase the explained variance (adjusted R2 = 0.28).
This study aimed to evaluate which individual factors that could predict the activity, disability, and pain intensity after 8 weeks of deadlift training in patients with mechanical low back pain. The main result was that measures of disability, pain intensity, and performance on the Biering-Sørensen test at the beginning of the training period could predict activity, disability, and pain intensity at 8 weeks' follow-up (i.e., after the training period). The higher the disability and pain intensity and the lower performance on the Biering-Sørensen test, the less likely that participants were to benefit from deadlift training. These findings suggest that pain intensity, disability, and performance on the Biering-Sørensen test should be considered before recommending an individual with low back pain to start deadlift training. A rough estimate of values corresponding to high pain intensity and low performance on the Biering-Sørensen test could be pain intensity above 60 mm on the visual analog scale (i.e., above the third quartile in this study) (Table 1) and an ability to hold the Biering-Sørensen test position for less than 60 seconds (i.e., below the first quartile in this study) (Table 1).
Baseline performance on the Biering-Sørensen test was a predictor in all 3 models, indicating that it can be a more robust predictive variable (32) than pain intensity and disability. The underlying mechanism of the relevance of the Biering-Sørensen test is arguably that it reflects an ability to activate the stabilizing hip and back extensors for a sustained period, which is also important in deadlift training. An inability to optimally activate the stabilizing hip and back extensors during deadlift training can lead to sustained tissue stress and ongoing pain (28). To increase muscle strength and avoid stress on passive tissues, the physical therapist instructed all participants to brace the trunk muscles to stabilize the lumbar spine during lifts. However, we cannot assure that all lifts for every participant were performed with optimal activation of the stabilizing hip and back extensors. Earlier research has shown an association between poor performance on the Biering-Sørensen test and high pain intensity (16,26). The fact that both the Biering-Sørensen and pain intensity were associated with activity, disability, and pain intensity at follow-up confirms this association in our study.
In the predictive models, baseline pain intensity was a predictor for activity and disability, which implies that participants with high pain intensity at baseline should not use the deadlift as a rehabilitative exercise. There could be several reasons for this. First, it has been proposed that the adaptation to pain in motor behavior involves a redistribution of muscle activity and changes in mechanical behavior (13). Participants with high pain intensity might therefore have had nonoptimal muscle activation patterns while performing the deadlift exercise. An altered activation due to pain might explain why pain intensity was associated with performance on the Biering-Sørensen test. Second, regarding the execution of the deadlift exercise, although the physical therapist instructed the participants to contract their trunk muscles to stabilize their lumbar spine in a neutral position, their deep stabilizing muscles might not have been optimally activated. Tsao and Hodges have shown that, to optimally retrain the feed-forward mechanism of the transversus abdominis muscle, isolated retraining, not general trunk strengthening with a sit-up exercise, is most effective for patients with recurrent low back pain (34). However, to ensure optimal performance regarding muscle activation during deadlift training, more advanced equipment (e.g., electromyography) would have been needed. Third, pain not only inhibits optimal muscle recruitment patterns but can also influence motor learning (13). More specifically, it has been suggested that pain may disturb motor learning due to its interference in the quality of performing a task that is being practiced (e.g., the deadlift exercise) (12). It might have therefore taken more time for participants with high pain intensity to learn how to perform the deadlift with the proper technique. Fourth, participants with high pain intensity might have needed treatment for pain alleviation, as suggested by Widerström et al. (36), before starting the deadlift training to enhance their potential to perform the exercise properly. Altogether, to benefit from deadlift training, participants with low endurance in the hip and back extensors and high pain intensity might need isolated retraining in activating local and global stabilizing muscles and/or pain alleviation before initiating training with the deadlift exercise.
Disability at baseline was included as a predictor in the predictive model for pain intensity at follow-up. The questionnaire used for measuring disability, the Roland-Morris Disability Questionnaire, is designed to measure disability during a fairly wide range of activities commonly impaired in patients with low back pain (e.g., bending or kneeling down) (15). However, it also addresses aspects of how the patients' mood and/or behavior are affected by their low back pain (15). Participants in this study were classified with a dominating pattern of mechanical low back pain, defined as pain that can be aggravated or relieved by different movements or postures (30). Because high pain intensity is reflected in many movements, postures, and behaviors, a correspondingly high score on the Roland-Morris Disability Questionnaire is natural. However, in accordance with previously mentioned consequences of high pain intensity, participants who scored high on the Roland-Morris Disability Questionnaire might have been impaired in such an extensive way in various activities demanding a wide range of body functions that they were simply unable to reap the benefits of the deadlift training.
In univariate analysis, the prone bridge test was significant (p ≤ 0.05) in predicting the follow-up value of activity, yet showed a low degree of explained variance (adjusted R2 = 0.09). Furthermore, the side bridge (i.e., right side) and movement control tests could not significantly explain any variance in the final predictive models. It therefore seems that these tests do not measure any dimensions that are important to predicting which patients can benefit from deadlift training. The results indicate that the prone bridge, side bridge, and movement control tests might not reflect the ability to control the lumbar neutral position during deadlift training to the same extent as the Biering-Sørensen test. Control of the neutral position of the lumbar spine during a flexion load and the strength and endurance of the hip and back extensors as measured with the Biering-Sørensen tests seem to be important aspects in deadlift training.
The baseline variables of age, sex, and body mass index were also unable to predict activity, disability, or pain intensity after deadlift training. The initial pilot study of the deadlift exercise (14) included physically active, middle-aged men. Our results show that women and men aged 26–60 years, with a dominating pattern of mechanical low back pain, can benefit from deadlift training, if not restricted by the significant predictors of disability and pain intensity or performance on the Biering-Sørensen test.
Generally, participants in the present study reported less disability and pain-related fear of movement and higher levels of activity than those with low back pain in previous studies (4,9,20), possibly because this study included only participants with a dominating pattern of mechanical low back pain. The low values for pain-related fear of movement might explain the nonsignificant influence of these independent variables due to floor effects in the models. Regarding activity, the Patient-Specific Functional Scale measured the degree of limitations in activities that each participant reported to be most affected by his or her low back pain. Deadlift training might thus benefit participants who report activities such as bending forward or lifting, but not activities that involve sitting or standing, which might explain why the baseline values of activity did not become a significant predictor in any of the predictive models.
Some important methodological considerations in this study should be discussed. The design of this study (i.e., a secondary analysis of a randomized controlled trial) is not optimal for creating predictive models. However, we chose this design to generate possible predictive variables for further prospective studies. Furthermore, the relatively small sample size (n = 35) could have affected the predictive models. According to Tabachnick and Fidell (33), regarding cases relative to each independent variable in regression analysis, a ratio of 20:1 is recommended to achieve valid results, although Field (10) recommends a ratio of 10:1 or 15:1. In the present study, all 3 predictive models had a ratio of 16:1, which clearly meets Field's (10) criteria, yet poses a slight power issue in relation to the recommendations of Tabachnick and Fidell (33). The models have adjusted R2 values ranging from 0.17 to 0.28. Compared with previous studies' ranges of 0.41–0.62 (3), 0.17–0.48 (2), and 0.15–0.55 (25), the explained variance in our models might be deemed somewhat weak. However, we consider that the predictive variables were found to be reasonable for and representative of the selected group of participants, although we might not have targeted all potential predictors. This topic could be an area for future studies.
We chose the follow-up values of the Patient-Specific Functional Scale, the Roland-Morris Disability Questionnaire, and a visual analog scale to represent activity, disability, and pain, respectively, at the end of the training period. We chose these domains because pain and disability are the most common reasons why individuals with low back pain seek care. Moreover, it has been suggested that different aspects of related outcome measures should be included in predictive modeling because such aspects reduce the chance of identifying single false-positive predictors (i.e., type 1 errors) (3).
In conclusion, the results of the analyses suggest that, in patients with mechanical low back pain, endurance of hip and back extensors (Biering-Sørensen test), pain intensity, and disability are important factors to assess before initiating deadlift training, if the goal in training is a high score in activity, low scores of disability and pain intensity. Further research should test the results presented in this study with a larger sample to examine whether the variables found in the present study can be confirmed as predictors.
Strength and conditioning professionals should not hesitate to use the deadlift exercise in their everyday practice, but before considering deadlift training for individuals with mechanical low back pain, our results suggest that pain intensity and the endurance of the hip and back extensors should be evaluated. For example, if low endurance of the hip and back extensors and high pain intensity are found in an individual with mechanical low back pain, then other interventions should be considered before initiating deadlift training. However, regardless of patients' age, sex, body mass index, pain-related fear of movement, movement control, and activity, the deadlift exercise seems to be an effective intervention. Finally, we stress the importance of strength and conditioning professionals' being skilled in technique instructions and providing feedback when instructing the deadlift exercise and progressing training for individuals with mechanical low back pain.
The authors express their gratitude to all participants of the study. This research was supported by 2 grants from Visare Norr and Norrbottens Läns Landsting. Its results do not constitute an endorsement of the product by the authors or of the National Strength and Conditioning Association. The authors declare they have no competing interests or professional relationships with the companies or manufacturers who might benefit from the results of the present study. This study was conducted in an outpatient physical therapy clinic in Umeå, Sweden.
1. Aasa B, Berglund L, Michaelson P, Aasa U. Individualized low load motor control
exercises and education vs a high load lifting exercise and education to increase function in patients with low back pain—a randomized controlled trial. J Orthop Sports Phys Ther 45: 77–85, 2015.
2. Beneciuk JM, Bishop MD, Fritz JM, Robinson ME, Asal NR, Nisenzon AN, George SZ. The STarT back screening tool and individual psychological measures: Evaluation of prognostic capabilities for low back pain clinical outcomes in outpatient physical therapy settings. Phys Ther 93: 321–333, 2013.
3. Cook CE, Learman KE, O'Halloran BJ, Showalter CR, Kabbaz VJ, Goode AP, Wright AA. Which prognostic factors for low back pain are generic predictors of outcome across a range of recovery domains? Phys Ther 93: 32–40, 2013.
4. Costa LO, Maher CG, Latimer J, Hodges PW, Herbert RD, Refshauge KM, McAuley JH, Jennings MD. Motor control
exercise for chronic low back pain: A randomized placebo-controlled trial. Phys Ther 89: 1275–1286, 2009.
5. Duehring MD, Feldmann CR, Ebben WP. Strength and conditioning practices of United States high school strength and conditioning coaches. J Strength Cond Res 23: 2188–2203, 2009.
6. Dworkin RH, Turk DC, Farrar JT, Haythornthwaite JA, Jensen MP, Katz NP, Kerns RD, Stucki G, Allen RR, Bellamy N, Carr DB, Chandler J, Cowan P, Dionne R, Galer BS, Hertz S, Jadad AR, Kramer LD, Manning DC, Martin S, McCormick CG, McDermott MP, McGrath P, Quessy S, Rappaport BA, Robbins W, Robinson JP, Rothman M, Royal MA, Simon L, Stauffer JW, Stein W, Tollett J, Wernicke J, Witter J. Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain 113: 9–19, 2005.
7. Ebben WP, Carroll RM, Simenz CJ. Strength and conditioning practices of National Hockey League strength and conditioning coaches. J Strength Cond Res 18: 889–897, 2004.
8. Escamilla RF, Francisco AC, Kayes AV, Speer KP, Moorman CT III. An electromyographic analysis of sumo and conventional style deadlifts. Med Sci Sports Exerc 34: 682–688, 2002.
9. Ferreira ML, Ferreira PH, Latimer J, Herbert RD, Hodges PW, Jennings MD, Maher CG, Refshauge KM. Comparison of general exercise, motor control
exercise and spinal manipulative therapy for chronic low back pain: A randomized trial. Pain 131: 31–37, 2007.
10. Field A. Discovering Statistics Using IBM SPSS Statistics. London: SAGE Publications Ltd, 2013.
11. Hicks GEFJ, Delitto A, McGill SM. Preliminary development of a clinical prediction
rule for determining which patients with low back pain will respond to a stabilization exercise program. Arch Phys Med Rehabil 86: 1753–1762, 2005.
12. Hodges PW. Pain and motor control
: From the laboratory to rehabilitation. J Electromyogr Kinesiol 21: 220–228, 2011.
13. Hodges PW, Tucker K. Moving differently in pain: A new theory to explain the adaptation to pain. Pain 152: S90–S98, 2011.
14. Holmberg D, Crantz H, Michaelson P. Treating persistent low back pain with deadlift training: a single subject experimental design with a 15-month follow-up. Adv Physiother 14: 61–70, 2012.
15. Johansson E, Lindberg P. Subacute and chronic low back pain. Reliability and validity of a Swedish version of the Roland and Morris Disability Questionnaire. Scand J Rehabil Med 30: 139–143, 1998.
16. Latimer J, Maher CG, Refshauge K, Colaco I. The reliability and validity of the Biering-Sorensen test in asymptomatic subjects and subjects reporting current or previous nonspecific low back pain. Spine 24: 2085–2089, 1999; discussion 2090.
17. Lundberg M, Styf J, Carlsson S. A psychometric evaluation of the Tampa scale for kinesiophobia—From a physiotherapeutic perspective. Physiother Theory Pract 20: 121–133, 2004.
18. Luomajoki H, Kool J, de Bruin ED, Airaksinen O. Reliability of movement control tests in the lumbar spine. BMC Musculoskelet Disord 8: 90, 2007.
19. Luomajoki H, Kool J, de Bruin ED, Airaksinen O. Movement control tests of the low back; evaluation of the difference between patients with low back pain and healthy controls. BMC Musculoskelet Disord 9: 170, 2008.
20. Macedo LG, Latimer J, Maher CG, Hodges PW, McAuley JH, Nicholas MK, Tonkin L, Stanton CJ, Stanton TR, Stafford R. Effect of motor control
exercises versus graded activity in patients with chronic nonspecific low back pain: A randomized controlled trial. Phys Ther 92: 363–377, 2012.
21. McGill SM, Childs A, Liebenson C. Endurance times for low back stabilization exercises: Clinical targets for testing and training from a normal database. Arch Phys Med Rehabil 80: 941–944, 1999.
22. Nuzzo JL, McCaulley GO, Cormie P, Cavill MJ, McBride JM. Trunk muscle activity during stability ball and free weight exercises. J Strength Cond Res 22: 95–102, 2008.
23. Ostelo RW, de Vet HC. Clinically important outcomes in low back pain. Best Practice Research Clin Rheumatology 19: 593–607, 2005.
24. Ostelo RW, Deyo RA, Stratford P, Waddell G, Croft P, Von Korff M, Bouter LM, de Vet HC. Interpreting change scores for pain and functional status in low back pain: Towards international consensus regarding minimal important change. Spine 33: 90–94, 2008.
25. Rasmussen-Barr E, Campello M, Arvidsson I, Nilsson-Wikmar L, Ang BO. Factors predicting clinical outcome 12 and 36 months after an exercise intervention for recurrent low-back pain. Disabil Rehabil 34: 136–144, 2012.
26. Ropponen A, Gibbons LE, Videman T, Battié MC. Isometric back extension endurance testing: Reasons for test termination. J Orthop Sports Phys Ther 35: 437–442, 2005.
27. Roussel NA, Nijs J, Mottram S, Van Moorsel A, Truijen S, Stassijns G. Altered lumbopelvic movement control but not generalized joint hypermobility is associated with increased injury in dancers. A prospective study. Man Therapy 14: 630–635, 2009.
28. Sahrmann SA. Diagnosis and Treatment of Movement Impairment Syndromes. St. Louis, MO: Mosby, 2002.
29. Schellenberg KL, Lang JM, Chan KM, Burnham RS. A clinical tool for office assessment of lumbar spine stabilization endurance: Prone and supine bridge maneuvers. Am J Phys Med Rehabil 86: 380–386, 2007.
30. Smart KM, Blake C, Staines A, Thacker M, Doody C. Mechanisms-based classifications of musculoskeletal pain: Part 3 of 3: Symptoms and signs of nociceptive pain in patients with low back (+/- leg) pain. Man Therapy 17: 352–357, 2012.
31. Stradford P, Westway M, Binkely J. Assessing disability and change on individual patients: A report of a patient specific measure. Physiother Can 47: 258–263, 1995.
32. Sun X, Briel M, Walter SD, Guyatt GH. Is a subgroup effect believable? Updating criteria to evaluate the credibility of subgroup analyses. BMJ 340: c117, 2010.
33. Tabachnick BG, Fidell LS. Using Multivariate Statistics, 6th ed. Boston: Pearson, 2013.
34. Tsao H, Hodges PW. Immediate changes in feedforward postural adjustments following voluntary motor training. Exp Brain Res 181: 537–546, 2007.
35. Vibe Fersum K, O'Sullivan P, Skouen JS, Smith A, Kvale A. Efficacy of classification-based cognitive functional therapy in patients with non-specific chronic low back pain: A randomized controlled trial. Eur J Pain 17: 916–928, 2013.
36. Widerstrom B, Olofsson N, Arvidsson I, Harms-Ringdahl K, Larsson UE. Inter-examiner reliability of a proposed decision-making treatment based classification system for low back pain patients. Man Therapy 17: 164–171, 2012.
Keywords:Copyright © 2015 by the National Strength & Conditioning Association.
motor control; resistance training; Biering-Sørensen test; pain intensity; prediction