1.1

Introduction

Clinical examination of the cervical spine is a prerequisite for the management of neck pain, on both the exploratory and therapeutic levels. However, this type of examination is far from having been standardized and varies according to the examiner’s training. Physicians, osteopaths, chiropractors and physiotherapists do not use the same manoeuvres and sometimes seek to measure very different data. Moreover, the agreement of these various manoeuvres is poorly documented in the literature; studies on this theme are very rare, often patchy and only consider one particular aspect of this problem. In fact, a highly reproducible investigational manoeuvre is prerequisite for assessment of its validity, that is to say, its true significance and utility. Here, we decided to study the interexaminer agreement for the cervical spine examination described by R. Maigne and modified by J.-Y. Maigne (such as we routinely apply in manual medicine) and to compare our results with those in the literature.

1.2

Materials and methods

The study population consisted of a consecutive series of patients suffering from acute or chronic common neck pain and attending an initial consultation in our hospital’s physical medicine department. The inclusion conditions were as follows. The pain over the previous 24 hours had to exceed a score of at least 4 cm on a 0 to 10 cm visual analogue scale (VAS). The pain had to be predominantly in the neck; it could radiate to the head, towards the upper back or towards the arm without, however, extending below the elbow. We excluded patients suffering from cervical radiculopathy with pain below the elbow, upper back pain, shoulder blade pain or a headache in the absence of associated neck pain. Likewise, we excluded subjects whose command of the French language was poor, subjects having suffered a workplace accident, those involved in a legal dispute and those having undergone neck surgery.

After having given their verbal agreement, patients were examined on inclusion and before a full, rigorous interview by one of two examiners (chosen in a predetermined, random order). The second examiner left the room during the initial examination and was not allowed to talk to the other examiner or discuss matters with the patient. In the second examination, the patient latter was instructed to give a “yes” or “no” answer during the pain provocation manoeuvres, without reference to the previous examination. The senior examiner had 25 years’ experience in manual medicine. The junior examiner was being trained in this discipline. The study lasted 12 months (from May 2006 to May 2007) and was preceded by a 15-patient pilot series (in order to homogenize the examination technique).

1.3

Results

Fifty-nine patients (33 women and 26 men) were included in the study. The mean ± S.D. age was 46.3 ± 12 years (range: 25 to 79).

1.4

Discussion

The present study was prompted by the quality criteria selected in the meta-analysis by Seffinger et al. . Our sample size was similar to the literature average and just largely exceeded the reference number of 30 subjects. Three potential sources of bias should be mentioned. The first is potential sensitization of patients after the first examination, with a possibly more painful second examination and/or erroneous responses. In fact, the total pain scores for the first and second examinations were comparable ( p = 0.8) and thus, there was no modification due to the dual examination. The second potential bias relates to the fact that one of the examiners was a physician undergoing training in manual medicine. His lack of experience could have altered the results. However, we first worked on a pilot series in order to correct and harmonise our examination techniques and 15 patients were probably enough for this purpose. Furthermore, the junior examiner’s lack of experience may even have made him more likely to objectively report what he had observed, without trying to minimize or exaggerate the results according to what would have been expected from theory. The third potential source of bias concerns the difficulty that patients had in simply answering “yes” or “no” to the question “does it hurt when I apply pressure here?” when the pain was mild. This difficulty is likely to reduce the size of the kappa coefficient. One solution would be to use a VAS to evaluate the pain provoked; however, Pool et al. have shown that the patients are not able to reliably quantify the same pain provoked in two examinations with a 15-minute interval .

There are few literature studies investigating interexaminer agreement for examination of the spine and cervical soft tissues and even fewer concerning segmental examination. Hubka and Phelan examined 30 patients level by level while looking for cervical spine tenderness but all the subjects had strictly unilateral, mechanical neck pain and were examined only on the spontaneously painful side – which would have somewhat improved the examination’s performance. These authors’ kappa of 0.68 (substantial agreement; confidence interval: 0.47–0.89) was slightly higher than ours (0.53, moderate agreement; confidence interval: 0.312–0.73). Like us, Van Suijlekom et al. classified the most sensitive level into three groups (the upper, median or lower cervical spine). Their kappa values ranged from 0.2 to 0.6 (fair to moderate agreement). Lastly, Cleland et al. followed a slightly different protocol because the segmental pain was identified by testing mobility vertebra per vertebra rather than by application of pressure . The kappa was below 0.27 (fair), except for C6 and C7 (0.55 and 0.90, respectively). There are only slightly more publications concerning the agreement of the examination of muscles or their insertions . The trapezius and levator scapulae muscles were examined but neither the splenius cervicis nor the entire length of the semispinalis muscles were assessed. These results are given in Table 4 . One can note that the degree of agreement ranges from fair to moderate. Our results are less good than those from Andersen and Gaardboe but their study mainly recruited fibromyalgia patients. There are a few more studies on the degree of agreement for cervical mobility measurement but the work was performed under different conditions, such as with a goniometer or by studying active mobility or segmentally (level by level). Youdas et al. noted poor agreement (by using intra-class correlation coefficients) when mobility was evaluated visually . In contrast, Hoppenbrouwers et al. found a kappa of 0.54 (moderate agreement), Fjellner et al.’s kappa values were between 0.41 and 0.6 (moderate agreement) and Pool et al.’s values ranged from 0.05 to 0.61 (from slight to substantial agreement) . Our method of calculation ( p value) does not facilitate comparison with other studies but we did obtain a kappa of 0.57 (moderate agreement) for determination of the side on which rotation was most restricted. There again, we noticed the two examiners’ difficulty in agreeing when the rotation difference was less than 10°. Regarding pain in flexion or extension, one can compare the literature values of 0.53 and 0.67 and 0.55 and 0.23 with our data of 0.71 and 0.76 (substantial agreement), respectively.

The overall interexaminer agreement for our clinical neck examination was moderate (with a kappa > 0.4) and encourages us to continue to use the procedure on a routine basis and to teach it. Only the rotational mobility measurement and the pain identification during flexion or extension showed a strong degree of agreement. This indicates that future studies on validation of this clinical examination should focus on the latter parameters. In fact, as emphasized by Bogduk and McGuirk, none of the manoeuvres that we evaluated here has been validated. The true significance of mobility anomalies or muscle tenderness is not known . It is also not known whether a positive segmental examination corresponds to a vertebral element (the facet joint, the disc or the periosteum) or to the overlying muscles.

We also made two other interesting findings. For the trapezius and the splenius cervicis muscles, the presence of abnormal muscle sensitivity did not correspond to a precise cervical level. In contrast, sensitivity of the levator scapulae muscle was associated with pain in the median or upper cervical spine but not its lower segment. The second finding relates to the presence of a correlation between the number of tender spots and the INDIC questionnaire score; the higher the number of tender spots, the greater the functional impairment. Two studies have already noted that the presence of a tender spot in palpation of the cervical facet joint enables relatively good differentiation between neck pain sufferers and healthy subjects when the two groups are formed on the basis of functional questionnaire scores . Our results are in line with this previous work but provide a new and significant element – that clinical examination alone enables the physician to judge the extent of the functional impairment.

1.5

Conclusion

Our routine neck examination displayed moderate interexaminer agreement, although this does not necessary imply that the technique is valid. The presence of pain at the insertion of the levator scapulae muscle is associated with moderate to intense neck pain. The number of tender spots reflects the functional impairment experienced by the patient.