Introduction

Soccer is the most widely played sport worldwide, with more than 265 million currently active players [1]. Head impacts are common in soccer and include player-to-player, player-to-goalpost and player-to-ground collisions (also known as unintentional head impacts) as well as “heading” where players intentionally use their unprotected head to deliberately direct the ball in play [2]. In addition to concussion, soccer players are at risk for potentially adverse effects of these repetitive head impacts (RHI). Multiple prior studies have shown that RHI is associated with adverse structural and functional brain effects, including lower white matter integrity, lower scores on neuropsychological testing and greater psychological distress [3–15]. These effects appear to be independent of the effects of unintentional head impacts and concussion [16, 17].

One of the main functions of the neck musculature is to provide stability for the head. A soccer player’s head velocity and acceleration during heading is limited by the contraction of neck musculature in reaction the impulse caused by contact of the ball with the head [18]. However, the more than 25 neck muscle pairs have complex lines of action and multiple points of insertion which make it challenging to study neck biomechanics [19]. The extent to which the neck resists change in position when a force is applied, is determined by active contraction of neck musculature and the passive resistance to compression by tendons, ligaments, and bone structure. Appropriate neuromuscular control of the relevant neck muscles is necessary to stabilize the head in the response to an impact [20, 21]. Neural control of neck musculature is dictated, in part, by the biomechanical constraints intrinsic to the individual [19].

Neck strength and anthropometrics control head and neck kinematics and may incorporate the torso as an effective mass [12]. The resulting strain on the neck and head can lead to a range of injuries following sport-related head impacts [22]. A study of youth and collegiate soccer players (n = 100) found that athletes with smaller head and neck size and lower neck strength were at risk for greater head acceleration [12]. Neck strength is hypothesized to be protective against head and neck injuries because stronger necks would better stabilize the head during an impact [23–25]. Studies on the protective role of greater neck strength in sports have shown mixed results. Neck strength was not found to be predictive of concussion risk for ice hockey players [26]. However, a small study (n = 17) of adolescent female soccer players found that greater neck strength is negatively correlated with head acceleration during heading [9], and a large study (n = 6,704) of male and female adolescent basketball, soccer, and lacrosse athletes found that for every 1 pound increase in neck strength, the odds of concussion decreased by 5% [27]. A recent systematic review by Elliott et al. found that higher neck strength may be associated with a lower risk of sport-related concussion [24]. While these latter studies identified a potential role for neck strength in prevention of recognized concussion, they did not address the role of RHI, which entail many impacts over time that do not lead to recognized concussion.

Understanding the potential for neck strength to mitigate adverse effects of RHI in sport could lead to identification of a modifiable risk factor [28, 29] for the previously described adverse effects of sports-related RHI [3, 30–32]. In addition, RHI from sports have been proposed to be a potential mechanism leading to the development of neurocognitive decline later in life [33, 34]. If neck strength has the potential to reduce the risk for adverse consequences of soccer heading, neck strengthening exercises could be a protective measure to reduce head acceleration and risk of adverse effects from RHI.

The purpose of this study was to test the hypothesis that neck strength or anthropometrics moderate previously reported adverse associations of both short-term (2 weeks) and long-term (12 months) soccer RHI exposure with neurocognitive performance, and with central nervous system (CNS) symptoms in adult amateur players [3, 30–32]. Based on the literature [8, 12], we hypothesize that neck strength (forward flexion, extension, left lateral flexion, right lateral flexion) and anthropometrics (head circumference, neck length, neck circumference and neck volume) moderate adverse associations of soccer RHI exposure with CNS symptoms.

Methods

The Einstein Soccer Study is a multifaceted longitudinal study of repetitive head impacts (RHI) and its consequences in adult amateur soccer players [30]. We analyzed prospectively collected data including assessments of RHI exposure, cognitive performance, and neck strength and anthropometrics. The same data collection and assessments were used for all participants.

This study was carried out in accordance with the recommendations of the Albert Einstein College of Medicine Institutional Review Board with written informed consent from all subjects.

Statistical power was based on published and preliminary data on the association of RHI with cognitive outcomes used in the Einstein Soccer Study cohort [35]. We utilized all available participant data without exclusion to minimize bias. Participants were 380 adult amateur soccer players who completed a total of 1,307 study visits (1–7 visits per participant over a two-year period in a longitudinal study of soccer play and its consequences). Soccer players were recruited between November 2013 and May 2018 by print and Internet advertisement and through soccer leagues and clubs in New York City and surrounding areas. Interested individuals were directed to an enrollment website, which, after informed consent, collected screening information. A research team member contacted qualifying individuals, confirmed eligibility, willingness to participate in the study and invited to enroll. Inclusion criteria were: aged 18–55 years old; at least 5 years of active amateur soccer play; current active amateur soccer play; at least 6 months of amateur soccer play annually; and English language fluency. Participants were asked to report neurological or other medical diagnoses. Exclusion criteria were: schizophrenia, bipolar disorder; current neurological disorder; pregnancy; and medical contraindication to MRI. Details of the overall study design have been published previously [17, 35].

Results

Participant characteristics and sex differences

Participant characteristics at baseline are shown in Table 1. Women and men differed greatly on nearly all anthropometric and strength measures, with women exhibiting lesser neck strength compared to males (Fig 1). To avoid confounding, subsequent analyses were stratified by sex.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Neck strength and anthropometric measures for men and women.

Welch Two Sample t-test: *p<0.05; **p<0.01; ***p<0.001.

https://doi.org/10.1371/journal.pone.0302463.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Participant characteristics.

https://doi.org/10.1371/journal.pone.0302463.t001

Tests of moderation by neck strength and anthropometrics

Results are shown in Table 2 for moderation by each of the three principal components of eight previously reported adverse associations, 7 of short- or long-term heading with measures of cognitive performance and one of unintentional head impacts with CNS symptoms (see Methods). Out of 48 tests, only PC1 for women modified the effect of 12mo-heading on ISL at a significance level that exceeded the Bonferroni-corrected significance threshold. The details for this regression model are shown in S2 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Moderating effect of neck strength or anthropometrics on associations of heading with outcome measures.

https://doi.org/10.1371/journal.pone.0302463.t002

Sensitivity analyses (S4 and S5 Tables) showed that exclusion of participants over 50 years did not alter the results, from which we infer that inclusion of these older individuals, who were part of the study cohort, did not bias the regression models.

Discussion

The purpose of this study was to investigate whether neck strength or anthropometrics moderate adverse associations of soccer heading or unintentional head impacts with cognitive performance or CNS symptoms among adult amateur players. Among male soccer players, neither neck strength nor anthropometrics modified associations of either soccer heading or unintentional head impacts with either cognitive performance or CNS symptoms. Among female soccer players, PC1 (representing neck strength) moderated the association of 12 month heading with ISL (Table 2). The direction of this association suggests that greater neck strength worsens the adverse association of heading with cognitive performance among female soccer players. No other moderating effect of either neck strength or anthropometrics with either cognitive performance or CNS symptoms met the Bonferroni-corrected significance threshold.

Prior studies have reported potential mechanisms that could explain the unique moderating effect we identified for women. Potential ways in which our findings could be understood in light of prior results include: (a) prior research has found women are more vulnerable to adverse effects of soccer heading on cognitive performance [53] and white matter microstructure [54]; (b) women with greater neck strength might perform heading drills during practice which have been associated with adverse effects [55]; (c) greater neck strength could reflect a style of play that alters risk, such as heading while standing vs jumping [56]; (d) suboptimal heading technique could be more prevalent among female players with greater neck strength, altering risk [57, 58]; or (e) greater neck strength among women might confer imbalanced neck stabilization, which has been found to lead to greater angular head acceleration during heading among collegiate players [59].

However, the statistical significance of this modifying effect was not robust to various sensitivity analyses we conducted, which included: (1) combined analysis of male and female players to increase sample size; (2) including both peak and average neck strength in the PCA; (3) imputing missing values; and (4) using an independent vs exchangeable correlation structure in the GEE.

In addition to the potential mechanistic explanations (a-e, above), we also note that since the sample size for female players is smaller than for men, this unique finding could be spurious and warrants replication in larger samples. We therefore conclude that our findings do not support mitigation by isometric neck strength alone of previously identified adverse associations of soccer heading with cognitive performance among either men or women.

Although stronger neck musculature may indeed stabilize the head and reduce head acceleration, the notion that strength alone mitigates deleterious cognitive effects of heading is not supported by our results. We also found that neck strength did not modify associations of unintentional head impacts with CNS symptoms. Although our findings in adult amateurs contrast with prior studies in pediatric cohorts, which suggested neck strength would protect against concussion from unintentional impacts [9, 27], these prior results may not be generalizable to adults who have larger head mass, increased neck girth/strength and more developed heading skills [60]. Additionally, neck strength may only confer benefit when muscle activation occurs at the time of impact [9, 14, 15, 61, 62], which is less likely to occur during an unanticipated impact.

In addition to our primary assessment of direct modifying effects of neck strength or anthropometrics, we explored two other areas related to neck strength among soccer players. First, we found, in line with previous studies, that anthropometrics are poor predictors of neck strength (S3 Table) [27, 39], suggesting that anthropometrics cannot serve as effective surrogates for direct neck strength measures. Second, we found heading to show little association with neck strength in both men and women, suggesting that high frequency of heading may not improve neck strength and that players with stronger necks may not necessarily head more often (S6 Table). As we did not find a modifying effect of isometric neck strength, these findings may seem insignificant. However, it has been shown in professional soccer players that higher heading exposure was associated with lower severity of symptoms of concussion, depression, anxiety, and sleep disorders [63]. While neck strength alone does not appear to modify the association of heading with cognitive performance, it may still moderate the effect of heading on subclinical structural brain features such as white matter microstructure [54] and could have relevance when viewed in the context of timing muscle activation to the moment of impact [9, 14, 15, 61, 62]. Indeed, without testing other aspects of neck function (e.g., timing of neck muscle contraction), we cannot be sure if the lack of a moderating effect by neck strength observed in our research is limited to the role of static isometric neck strength or could indicate that neck function overall does not mitigate effects of RHI. These remain important areas for future investigation.

Many prior studies supporting the role of neck strength in mitigation of adverse effects of RHI are limited by relatively small sample sizes (often fewer than 30 participants) [9, 64], limited ecological validity of the laboratory paradigms employed [20] or lack of control for relevant factors such as previous concussion [27]. Moreover, most of these studies reported on head acceleration, not functional outcomes, such as incidence of concussion, symptoms, or cognitive/structural effects [11, 65]. Thus, it remains an open question whether the biomechanical effect of neck strength on head acceleration, demonstrated in prior studies [8, 9, 65, 66], prevents injury or dysfunction [24, 62]. Our study addresses many of these gaps in knowledge, with a large sample size, observational design assessing real-world soccer play and cognitive performance, and statistical modeling that accounts for potential confounders.

Confirmation that neck strength stabilizes the head during sport-related head impacts and reduces the risk of brain injury in the context of soccer RHI, could set the stage for the use of neck strengthening as a protective feature in soccer [23, 66]. Mansell et al. found that neck strength training in collegiate male and female soccer players did not affect head acceleration during head impacts [14]. Research into neck strength training in managing the risk to concussion suggests that neck stiffness at the time of impact, rather than neck strength, may reduce the risk of concussion [29, 62, 67, 68]. Some have found that incorporating neck exercises results in a decrease in head acceleration by improving neuromuscular control [66, 69]. Whereas other have demonstrated that there is no reduction of head acceleration after isotonic neck exercises [70, 71]. Although there is limited literature on neck training to reduce head injury risk [72–74] or neurological changes [7, 75], no study to date has directly examined the effect of neck strength on the adverse effects of RHI, including heading.

The biomechanics of heading and head kinematics following impact are complex and are likely to vary not only due to neck strength but also with the timing of cervical muscle activation [20], and with heading technique, such as head-torso alignment and follow-through [60]. The benefits of neck strength might only be relevant when timing of muscle activation is synchronous with the moment of head impact [8, 24, 62]. Further research is warranted to determine if integration of strengthening with optimized aspects of heading technique could realize a significant protective effect.

Limitations

Our findings must be considered in light of several limitations. Fixed-frame dynamometry is an established method for measurement of neck strength. However, it largely reflects isometric strength, which might not fully reflect the true biomechanics of head impact. Some papers have argued for studying muscle strength under dynamic conditions (e.g., isokinetic) using EMG, hypothesizing that the pattern and timing of muscle activation may be more important than absolute strength [29]. It is unclear whether neck strength imbalance, which could potentially worsen heading-related effects [10], is better assessed by the difference or ratio of flexor and extensor muscle strength. We chose to calculate the ratio of forward flexion to extension as an estimate of neck strength imbalance, which is the approach commonly reported in the literature [25, 42–44]. We performed a sensitivity analysis, which showed the two approaches made minimal differences in our model outcomes and did not alter significance. Our study might not have been sufficiently powered to detect a moderation effect, leading to false negative inferences. However, we studied a larger sample (380 participants, 266 men and 114 women, who completed a total of 1307 visits) than prior studies, which did identify significant adverse associations of heading with cognitive performance [5, 6, 30, 35]. Thus, we expect to be well-positioned to detect a moderation effect of next strength on these associations. The heading exposure variable we employ is derived from self-report, which may be subject to bias. However, the instruments we employ have been extensively validated [36, 37] and elicit similar exposure-response relationships across independent samples of players. The cognitive effects we attribute to heading could in part be due to unintentional head impacts. However, we previously reported in this cohort that heading, not unintentional impacts, explains variance in cognitive performance measures [17]. To confirm this, we performed a sensitivity analysis including unintentional impacts as a covariate, which did not change our results. One possibility is that unintentional impacts are comparably rare (median four per year), compared to heading (median 678 per year) in our sample, which may be insufficient exposure to affect cognitive performance. Finally, our results cannot explicitly be generalized beyond the adult amateur league players in this study. Other groups, especially collegiate or professional players could respond differently if they have much greater neck strength.

Conclusion

Although prior research proposed isometric neck strength as a modifiable risk factor for brain injury due to soccer heading, our findings suggest that neither isometric neck strength nor head/neck anthropometric measures mitigate previously reported adverse associations of soccer heading with cognitive performance in adult amateur players. In addition, neither neck strength nor anthropometric measures mitigated previously reported associations of unintentional head impacts with CNS symptoms. Nonetheless, a more nuanced approach, which considers neck strength along with timing of muscle activation and optimal technique, might be necessary to identify the factors that can confer risk reduction. Future research should examine whether muscle strength under dynamic conditions moderates the association of repetitive head impacts with cognitive performance, CNS symptoms, and subclinical structural brain features of injury such as white matter microstructure.

Supporting information