Study identification and selection
The search produced 118 papers. The references were exported to reference manager software (EndNoteTM X9, Clarivate Analytics, Philadelphia, PA, USA). Duplicates (48 references) were removed either manually or automatically. The 70 remaining articles were screened for their relevance by reading the titles and the abstracts. This process resulted in the removal of 40 studies. Following the screening procedure, 30 articles were selected for in=depth reading and analysis. After reading the full texts, 10 studies were excluded because they did not focus on heading only (but involved also upper trunk mechanics), were not written in English (n= 4), had < 6 subjects (n = 2), and was performed with subjects < 14 (2). All 12 papers included in the review were biomechanical and considered the outcomes. The PRISMA flowchart is presented in Figure 1.
Twelve studies met the inclusion criteria. We classified AXIS quality scores according to the number of “Yes” responses for the 20 items for each study: 80% “yes” responses indicated high quality, 60%–80% indicated moderate quality, and < 60 % indicated low quality. Among the studies, 4/12(36.6%) had high quality, 5/11 (36.6%) had moderate quality, and 3 (27.7%) had low quality. Items related specifically to reporting quality had high scores; detail related to study design and possible biases were lower and more variable (Fig. 2).
Study characteristics and qualitative synthesis.
The characteristics of the heading studies included in the review are reported in Table 1.
Table 1. Summary of studies characteristics and outcomes.
Author
|
Aim of the study
|
Type of study
|
mean age
|
weight
|
height
|
subject number
|
qualification level
|
Sensor type
|
parameter measured
|
PLA
|
RA
|
RV
|
|
|
|
|
|
|
|
|
|
|
(g)
|
(rad/sec)
|
(rad/sec)
|
Filben et al. 1
|
categorize head impacts
|
O
|
19
|
62.3
|
170
|
16
|
NCAA div 1
|
Mouth
|
PLA, PRA,PRV
|
|
|
|
|
|
|
|
|
|
|
|
|
corner
|
22.9
|
2189.3
|
9.87
|
|
|
|
|
|
|
|
|
|
goal
|
24.3
|
2658.9
|
10.1
|
|
|
|
|
|
|
|
|
|
free
|
18
|
1843.3
|
8.43
|
|
|
|
|
|
|
|
|
|
live
|
18.8
|
1769.7
|
8.09
|
Reynolds et al.
|
quantify head impact
|
O
|
20.43
|
|
|
7
|
NCAA div 1
|
Mouth
|
PLA,PRA
|
19.1
|
3687
|
|
Lynall et al.
|
frequency and
|
O
|
19.1
|
63.7
|
168
|
22
|
NCAA
|
Mouth
|
PLA,PRA
|
12.51
|
2093
|
|
|
magnitude of impact
|
|
|
|
|
|
|
|
|
|
|
|
Filben et al. 2
|
difference younger/experienced
|
E
|
15.27
|
61.8
|
|
6
|
U15
|
Mouth
|
PLA,PRA,RV
|
11.7
|
9450
|
5.48
|
|
|
E
|
20.19
|
63.08
|
|
13
|
NCAA div 1
|
|
|
17.3
|
15600
|
7.63
|
Caccese et. Al 2
|
different heading scenarios
|
|
19.6
|
60.3
|
167.8
|
16
|
NCAA div 1
|
Head
|
PLA, RA
|
28.2
|
7100
|
|
Gutierrez et al.
|
Effect on neurocognitive function
|
E
|
15.9
|
59
|
165
|
17
|
HIGH SCHOOL
|
Head
|
FA
|
5.83
|
|
|
|
|
|
|
|
|
|
|
|
LA
|
6.96
|
|
|
|
|
|
|
|
|
|
|
|
RA
|
6.27
|
|
|
Nowak et al.
|
reliability of concussion
|
E
|
19.7
|
62
|
168.3
|
23
|
|
Head
|
PLA
|
16.1
|
|
|
|
assessment
|
|
|
|
|
|
>5 yrs
|
|
|
|
|
|
Tierney et al.
|
headgear effect of PLA
|
E
|
19.5
|
63.2
|
164
|
29
|
>5 yrs
|
Mouth
|
PLA
|
21.52
|
|
|
Muller et al.
|
asses neck and trunk
|
E
|
16.5
|
56
|
168
|
7
|
HIGH SCHOOL
|
Head
|
PLA
|
10.9
|
|
|
|
strenghtening
|
|
|
|
|
|
|
|
|
|
|
|
Kenny et al.
|
which kick produced more impact
|
E
|
19.9
|
65.5
|
169.3
|
13
|
VARSITY
|
Mouth
|
PLA, RA
|
22.2
|
2296
|
|
Mihalik et al.
|
characterize the impact
|
E
|
19.8
|
64
|
168.2
|
34
|
NCAA div 1
|
Mouth
|
PLA, RA
|
19
|
3457
|
|
Bretzin et a.
|
correlate PLA and RV
|
E
|
20.25
|
66.9
|
158.7
|
8
|
NCAA div 1
|
Head
|
PLA,RV
|
24
|
1416
|
|
|
and muscle strength
|
|
|
|
|
|
|
|
|
|
|
|
Eleven of the retrieved studies were published after 2017, indicating increased research interest in women’s safety in football after that year. Table 1 shows the summary of the main evidence found in each review. Most studies used single groups or parallel, non-randomized groups. Seven studies employed a mouthpiece (inside the mouth) sensor, and five studies used a head sensor (top of the head, or facial).
The 2 types of employed sensors differ greatly each other. Mouth sensors has been showed to be more precise than head-mounted sensors and to have a tighter skull-coupling, given a substantial low error in measuring linear and rotational acceleration 26, however, also mouthguard mounted sensors has been show to produce substantial errors in the measurement of angular acceleration of head 27. but must be noted than heading is a complex movement involving multiple joints, crossing different planes of motion, while sensors are sensitive to 3 axis of motion.
Three studies were observational, and nine were experimental.
Of the included articles, 1 study compared the magnitude of impact between practice and match setting, 1 study evaluated the differences between male and females, 2 studies characterized heading in relation to different play scenarios, 1 study identified differences in impact between young and experience players, 3 studies determined the effect of measured parameters of injury and neurocognitive functions and the relationship between impact characteristics and the likelihood of developing neurocognitive deficits, 2 studies correlated neck muscle strength and magnitude of the impact, 2 studies identified which kick produced higher impact in a heading response, and 1 study assessed the dumping effect of a headgear.
The mean sample size was 16.2±8.7 subjects (range 7–34). The average age, weight, and height of the subjects were 18.8±1.7 years, 62.3±2.9 kg, and 167.1±3.27 cm (BMI 23.3±2.3), respectively. One study did not report weight and height. Seven studies were performed in NCAA div 1 players, 2 studies were conducted in high school players, 1 study considered > 5 years of experience, 1 study focused on varsity, and 1 study was under 15 teams.
The measured impact parameters were as follows: peak linear acceleration (PLA; m/sec 2), mean rotational acceleration (RA; m/sec 2), and mean rotational velocity (RV; m/sec). One study reported frontal acceleration and lateral acceleration (FA; and LA m/sec2). Figure 3 report a graphical description of the 3 accelerations.
The mean values for PLA, RV, and RA were 17.61±6.27, 2584.6±1628.41, and 8.27±1.68). Six studies reported the frequency of heading but used 5 different metrics (per training session, per athletes ‘exposure time, in training and in game, and in mixed situations). Exposure to heading ranged from 2 to 10 per match/training unit. The measurement devices were essentially of two types: head (top of the head or face) and mouth (inside the mouth) sensors. We can summarize the main outcomes as follows. First, goal heading and response to kicking from long distances produced the highest rotational acceleration in comparison with other game scenarios 28, being the game situations influencing the heading kinematic 29. The reviewed studies didn’t found any difference between males and females in heading kinematics 30 and the magnitude of heading was higher during practice than during the game 31 . Experienced players are able to produce higher acceleration and velocities 32. Is not clear if medium intensity soccer heading significantly impair neurocognitive function: some studies showed an impairment of neurocognitive function in females soccer players on the short and long term 34,35,36, while other studies not 36. Albeit in one study were found very high RA (4509 to 8869 rad/sec -2), exceeding the recomme,35nded limits, no symptoms of concussion were observed 35. Wearing a soft-cell helmet increased the PLA 37. Head mass and neck girth were correlated negatively with PLA and RV 38. Increased strength in the neck muscles mitigated PLA 39.The heading response to a long kick produced a higher PLA 40,41.