Lumbar position sense and the risk of low back injuries in college athletes: a prospective cohort study

Intact proprioception is essential for movement control [1‐3]. In the spine, proprioceptive information is provided by structures present in the spinal ligaments, facet joints, intervertebral discs [4‐6], and paraspinal muscles [7, 8]. Muscle spindle density is high in deep paraspinal rotators, which are small muscles spanning one or two segments of the spine [9]. It is believed that the spindles in these muscles act as kinesthetic sensors that monitor trunk position and movement. It is these muscle receptors that are more likely responsible for information in the midrange of trunk motions [8, 10]. While joint receptors cannot be discounted, these structures are thought to provide more input toward the end range of joint positions [10]. However, altered joint afferent information can alter muscle activation [11]. Consequently proprioceptive information from both muscle and joint receptors may be an important aspect of trunk control of motion. Since the overwhelming majority of low back injuries (LBI) in athletes are classified as soft tissue injury, the mechanoreceptors embedded in these tissues could be involved [12‐14].

Proprioceptive impairments, reflected by poor joint position sense (PS), have been identified in numerous soft tissue injuries commonly suffered by athletes: anterior cruciate ligament deficiency [15, 16], ankle sprains [17], glenohumeral instability [18, 19], neck injury [20, 21] and low back pain (LBP) [8, 22‐27]. However, in the LBP literature, the evidence regarding the presence of proprioceptive impairments is not unanimous. Newcomer et al. (2000) [28] reported significantly larger repositioning error in patients with LBP in trunk flexion and significantly lower error in trunk extension when compared to a control group. Field et al. (1991) [29] found less variability in repositioning error in their LBP group and Parkhurst et al. found no correlation between directly measured proprioceptive variables and LBI, but instead reported its association with the asymmetry indices derived from these variables [30]. Finally, several studies demonstrated no proprioceptive impairments in individuals reporting LBP or injury [31‐34]. Differences in test conditions (body position, planes of motion, whether or not vestibular system is involved, lower body constraint), and subject characteristics could explain some of the divergent results in the literature.

Despite the above uncertainties, widely reported deficits in postural control [35‐39] and altered patterns of muscle response to sudden trunk loading [40‐42] among patients with LBP are hypothesized to be, at least in part, the result of injury to mechanoreceptors embedded in the soft tissues surrounding the lumbar spine. However, an alternative hypothesis would be that impaired spinal proprioception is a pre-existing risk factor that predisposes individuals to LBI.

The aims of this prospective study were to 1) examine whether differences in trunk PS exist between individuals with and without a history of LBI, and 2) determine if impairment in trunk PS results from injury or alternatively predisposes athletes to future LBI. Knowledge gained regarding trunk proprioceptive deficits would assist in developing targeted interventions for athletes. We hypothesized that athletes with a history of LBI would demonstrate less accurate and precise trunk repositioning and higher motion perception thresholds. Additionally, athletes with poor trunk PS would have a higher risk of sustaining a LBI than athletes with more accurate and precise trunk PS.

Our results reflected 292 college athletes who could be undoubtedly classified as injured or not injured based on the availability of records and our definition of injury. Their characteristics are presented in Table 1. At the start of the study, 60 athletes (21%) had a history of LBI within the last five years. Majority of them (43(72%)) sustained only a single LBI episode whilst the remainder had multiple episodes. During the follow-up period, 31 athletes (11%) became injured (Table 1). Of these, 12 athletes (39%) had a history of LBI. The athletes injured during follow-up were significantly (p < 0.01) taller (1.80(0.09) m vs. 1.76(0.10) m) and heavier (78.9(15.3) kg vs. 72.0(12.4) kg) than the uninjured athletes. The effects of history and subject characteristics on LBI were addressed in a previous publication [46]. The current report focuses solely on trunk PS data.

All data met assumptions of normality (Anderson-Darling test, Minitab, Inc.). The initial MANOVA returned no significant differences in trunk repositioning error in the transverse plane between the athletes with and without a history of LBI (p = 0.25) or between those who did and did not sustain a LBI during the follow-up period (p = 0.63) (Table 2). However, significant effects of test mode (passive or active) (p < 0.01) and gender (p = 0.04) were present. The post-hoc univariate analyses revealed that the athletes were significantly more accurate (AE, p < 0.01) and precise (VE, p < 0.01) in the active trunk repositioning tests as compared to the passive tests (Figure 2). Males were slightly (0.15°), but significantly less accurate (AE, p = 0.02) and less precise (VE, p = 0.01) than females (Figure 3).

There were no significant effects of any of the factors on MPT. On average, all athletes perceived their trunk rotation at 1.1° (SD = 1.0°).

The within-session reproducibility of the active and passive repositioning tests was good (0.47 < ICC < 0.61, 0.57° < SEM < 0.73°, Table 3). The reproducibility of MPT was excellent with ICC = 0.89 and SEM = 0.34° (Table 3).

There is inconsistency in the literature with regards to impairment in trunk PS and LBP. Some studies have found impairment [8, 22‐27], whilst others have not [31‐34]. Because of these inconsistencies, it was not possible to state objectively whether LBP was associated with impairment in trunk PS. To address this problem, we designed a large prospective study with a homogenous subject population using a similar protocol to that of Taimela, Leinonen, and colleagues [25‐27]. This protocol has the advantage in that it isolates proprioception to trunk sensory receptors, while other inputs from lower extremities, vision, and the vestibular system are removed. Given the results of this study, we would conclude that impaired trunk PS is not associated with LBP. It appears that the majority of back injuries in athletes do not involve significant disruption of trunk PS, nor does it appear that poor trunk PS predisposes athletes to LBI. In comparison to previous studies, our results are strengthened by the use of a large homogenous subject group, measurement of several aspects of trunk proprioception, isolation of the trunk from lower extremity input and standardization of the range of movement of each subject around the neutral trunk position.

It is possible that our athlete population differs from the general population used in others studies with positive findings. Perhaps, factors such as age or fitness levels can account for the differences between the LBP and healthy controls. Proprioception declines with age [47, 48] and more fit individuals may have more accurate joint PS. This notion has been supported by research demonstrating better knee MPT in trained gymnasts versus healthy non gymnasts [49].

It is also possible that other planes of motion can be affected more by LBP than the transverse plane used in our study. However, it should be noted that a number of the positive studies documented impairment in this plane [25‐27]. So it would be expected that if impairment exists, it would also be found in the transverse plane of motion.

If impairment in trunk PS was strongly related to LBP, the findings in the literature would be more consistent. O'Sullivan et al.[23] suggested that the non-homogeneity of patients in terms of their specific pathologies may be responsible for the conflicting findings in research on trunk PS and LBP. The majority of injuries suffered by the athletes in our study were classified by health professionals in general terms as sprains or strains, and we did not attempt to diagnose these injuries further. It could be that impairment in trunk PS is specific to a particular patient population and/or pathology. This is still a possibility that needs to be investigated further.

It is unlikely that measurement limitations in our study could be responsible for the lack of differences in trunk PS between the athletes with and without a history of LBI or those who did and did not get injured during the follow-up. Our measurement errors (SEM) varied from 0.34° to 0.73° for the various test modes, and are in line with similar studies reporting diminished trunk PS in the LBP populations [25, 30]. The magnitude of repositioning errors and the MPT obtained in the present study was also compatible with control groups in previous research (between 0.8° and 1.6°) [25, 30]. More importantly, however, since our method was sufficiently sensitive to detect the differences in trunk repositioning accuracy between the active and passive testing modes (to be discussed shortly), it is likely that our data truly reflect the lack of impairment in athletes with a history of LBI. With similar confidence in our prospective study design, we can also reject the hypothesis that impaired trunk PS is a risk factor for sustaining a LBI in athletes.

Results from our study suggest differences between active and passive testing modes. Similarly to most of the relevant literature [50‐56], we too found that active trunk repositioning resulted in smaller errors than passive trunk repositioning. It is generally agreed that afferent input from muscle spindle, encoding information tied to active movement, is in part responsible for a more accurate and precise joint PS in the active testing mode. However, other mechanisms, such as central corollary discharge, can be also used in a feedforward mechanism to assist in the reproduction of joint position [54].

Data from the current study revealed small gender differences in favor of females having slightly more accurate and precise trunk PS. However, clinical significance of differences in trunk repositioning error smaller than 0.15° is probably negligible. Thus, these findings should be interpreted more in line with other literature, which reported no gender differences [25, 57].

Many athletic rehabilitation programs emphasize proprioception training as it is believed that impaired joint PS may be a major risk factor for recurrent injuries [2, 58]. While it is true that a history of LBI was the single best predictor of future LBI in athletes [14], based on our data, the mechanism mediating such injuries is not likely an impairment in trunk PS. Even in the present study, the athletes with a history of LBI had a 3-times greater risk of sustaining a LBI during the follow-up [46], but their trunk PS was not different from the athletes who had no history of LBI or those who did not sustain a LBI during the follow-up. If impairments in trunk PS exist during acute stages of LBI, they appear to recover relatively rapidly and do not constitute a risk factor for recurrent LBI.