Diagnostic Value of History Taking and Physical Examination to Assess Effusion of the Knee in Traumatic Knee Patients in General Practice

Article Outline

• Abstract
• Methods
  • Patients
  • Data Collection
  • Statistical Analysis
• Results
  • Study Population
  • Magnetic Resonance Imaging Results
  • History Taking and Physical Examination
  • Diagnostic Value of History Taking and Physical Examination
• Discussion
• Conclusions
• Appendix
• References
• Copyright

Abstract 

Kastelein M, Luijsterburg PA, Wagemakers HP, Bansraj SC, Berger MY, Koes BW, Bierma-Zeinstra SM. Diagnostic value of history taking and physical examination to assess effusion of the knee in traumatic knee patients in general practice.

Objective

To assess the diagnostic value of history taking and physical examination for knee joint effusion in patients with a knee injury who consult their general practitioner (GP). In addition, to determine the association between effusion seen on magnetic resonance imaging (MRI) and internal derangement of the knee.

Design

Prospective, observational cohort study.

Setting

Primary care.

Participants

Patients (N=134) aged 18 to 65 years with a traumatic knee injury who consulted their GP.

Interventions

Not applicable.

Main Outcome Measures

Patients filled out a questionnaire, underwent a standardized physical examination and underwent an MRI scan to assess the presence of effusion. Multivariate logistic regression analysis was used to determine the diagnostic value of history taking and physical examination (P<0.10) as assessed by sensitivity, specificity, predictive values, and likelihood ratios. The relationship between effusion and internal derangement of the knee was assessed with a chi-square test.

Results

Of the 134 participating patients, 42 had knee joint effusion seen on MRI. Multivariate analysis showed an association with knee joint effusion for the symptom “self-noticed swelling” (history taking) and for the “ballottement test” (physical examination). The likelihood ratio positive (LR+) was 1.5 for self-noticed swelling and 1.6 for the ballottement test. These 2 combined improved the diagnostic value to an LR+ of 3.6. Effusion showed a positive association with internal derangement of the knee (chi-square 9.5); 31 of the 42 patients with knee joint effusion had internal derangement of the knee.

Conclusions

In patients with traumatic knee injury, knee joint effusion is frequently seen on MRI. The combination of self-noticed swelling and the ballottement test was of diagnostic value. Knee joint effusion was associated with internal derangement of the knee.

Key Words: History taking, medical, Knee injury, Magnetic resonance imaging, Physical examination, Primary care, Rehabilitation

List of Abbreviations: CI, confidence interval, GP, general practitioner, LR+, likelihood ratio positive, LR−, likelihood ratio negative, MRI, magnetic resonance imaging, OR, odds ratio, PVN, predictive value – negative, PVP, predictive value – positive

KNEE INJURY IS A FREQUENT reason to consult the GP. The incidence of traumatic knee injuries (excluding fractures) reported in Dutch general practice is about 5.3 per 1000 patients per year.¹ Effusion is frequently seen in traumatic knee patients and in secondary care a positive correlation between effusion (MRI) and internal derangement of the knee is reported.² Additionally, Kolman et al³ reported that symptomatic patients with no significant effusion in the lateral aspect of the suprapatellar pouch on MRI are free of internal derangement in 86% of the cases. In the Netherlands, about 25% of all patients presenting in general practice with traumatic knee injuries are referred to secondary care.⁴ Therefore, in traumatic knee patients it would be useful if the GP could detect knee joint effusion by means of history taking and physical examination.

Various questions from history taking (eg, Is your knee swollen?) and various physical examination tests are available to evaluate the presence and amount of knee joint effusion. Tests from physical examination include palpation of effusion and/or Baker cyst in the fossa poplitea,⁵ the peripatellar fluctuation test (minor effusion test),⁵ and the floating patella test (ballottement test).⁵, ⁶ However, the diagnostic value of history taking and physical examination is often questioned by clinicians.⁷, ⁸ Further, in routine practice it is still unclear what the diagnostic value of history taking and physical examination is, when and which test to use, and which test is most effective to evaluate effusion.

The present study aims to determine the diagnostic value of items from history taking and physical examination in detecting effusion of the knee joint as seen on MRI in traumatic knee patients in primary care. The second aim is to investigate the association between effusion seen on MRI and internal derangement of the knee in this primary care population. MRI was selected as the reference test because it is a highly accurate diagnostic tool for detecting effusion of the knee joint.³, ⁹

Back to Article Outline

Methods 

Patients 

The present study was performed within the HONEUR knee cohort¹⁰ (a prospective, observational cohort study of patients consulting the GP for all types of new knee complaints). In the present study, patients who had a traumatic knee complaint less than 5 weeks before consultation and were aged 18 to 65 years, but had no MRI contraindications (pregnancy, metal implants, or a pacemaker), were asked for informed consent for an additional MRI.

Data Collection 

An appointment was made for the MRI and patients filled out a baseline questionnaire¹⁰ which they had to bring along at the MRI examination. A standardized physical examination¹⁰ was performed immediately after the MRI by a trained physical therapist. The physical therapist was blinded for the MRI results, as was the radiologist who scored the MRI for the results of the physical examination and questionnaire. Both the patient and the GP were not informed about the outcome of the MRI or physical examination; this was to avoid influencing the behavior of the patient, or the management by the GP during follow-up.

The baseline questionnaire collected data on age, sex, socio-economic status, history of previous knee injuries and/or operations, present symptoms, mechanism of injury, level of activity in work, household, study and sports, pain severity, and the Lysholm knee score.¹¹ Physical examination of both knees consisted of inspection, palpation, assessment of effusion, passive range of motion in flexion and extension, knee stability tests, and meniscal tests. In the present study we focused on the question: did the patient notice any swelling of the knee and on the following tests (see Appendix 1): palpation of effusion or Baker cyst in the fossa poplitea, the peripatellar fluctuation test (minor effusion test), and the floating patella test (ballottement test).⁵, ⁶

MRI was scheduled 3 to 6 weeks after the initial trauma and was performed with a 1.0 Tesla device.^a Two radiologists classified the MRI scans independently from one another. Detailed information about the MRI procedure is reported elsewhere.¹²

Regarding effusion, patient outcome was defined as absent, small (fluid in 1 or 2 compartments), moderate (fluid in 3 compartments), or severe (fluid in 3 compartments with bulging of the capsule and pericapsular soft tissue) knee joint effusion (accumulation of intra-articular fluid in suprapatellar, medial or lateral compartment) as seen on MRI.¹² In this study we dichotomized effusion into none to small versus moderate to severe effusion. We designated moderate to severe effusion as being clinically important effusion.

The study protocol was approved by the Medical Ethics committees of the Erasmus Medical Center Rotterdam and of the Medical Center Rijnmond Zuid.

Statistical Analysis 

Descriptive statistics were used to present the results of the MRI. Univariate logistic regression analysis^b was used to determine the association between knee joint effusion and separate symptoms from history taking and signs from physical examination, expressed as OR. The symptoms and signs indicating a univariate association (P<0.10) with knee joint effusion were analyzed with multivariate logistic regression analysis (Backward Wald method, entry 0.10, removal 0.20).

We determined the diagnostic value of the isolated symptoms from history taking and signs from physical examination by calculating the sensitivity, specificity, PVP and PVN.¹³ We also determined the likelihood ratio for positive and negative examination.¹³ In general, an LR+ of 1 to 2 or an LR− between 0.5 and 1 changes the probability of the presence or absence of knee joint effusion by only a small degree.¹⁴ An LR+ of 2 to 10 or an LR− between 0.1 and 0.5 may be considered clinically important.¹⁴ An LR+ greater than 10 or an LR− less than 0.1 may have substantial impact on the probability of the diagnosis. Finally, we combined symptoms and signs from the composed diagnostic model for knee joint effusion (P<0.10) and determined the diagnostic value of these combinations.

The relationship between effusion and internal derangement of the knee (cruciate ligament lesion and/or meniscal lesion with the exception of degenerative lesions) was assessed with a chi-square test with a significance level of 0.05.

Back to Article Outline

Results 

Study Population 

Of the 184 eligible patients, 134 (73%) were included in the present study. No significant differences were found between the baseline characteristics of the participants and non-participants (data not shown). Table 1 presents the characteristics of the participating patients.

Table 1. Characteristics of the Participants (N=134) and Diagnosis as Seen on MRI

	n (%)
Characteristic
Mean age ± SD (y)	40.2±12.2
Sex, men	74(55.2)
Onset during sports activity	61(45.5)
Symptom side right	70(52.2)
Pain severity (0–10), mean ± SD	4.7±2.4
Lysholm knee function score (0–100), mean ± SD	63.7±18.9
Diagnosis as seen on MRI
No lesion or effusion⁎	14(10.4)
Contusion (effusion⁎ with no ligament or meniscal lesion)	38(28.4)
Clinically important effusion†	42(31.3)
Medial collateral ligament lesion	35(26.1)
Lateral collateral ligament lesion	8(6.0)
Anterior cruciate ligament lesion	28(20.9)
Posterior cruciate ligament	6(4.4)
Meniscal tear	47(35.1)

NOTE. Values are n (%) unless otherwise noted.

Of the 134 participating patients, 61 (45.5%) reported that the knee injury was caused by sport activities. Mean age ± SD of the participants was 40.2±12.2 years and 55.2% were men.

Magnetic Resonance Imaging Results 

Table 1 also gives the MRI results. The average period between trauma and MRI was 38 (range, 9–81) days; 70% of all patients had their MRI within 6 weeks after the initial trauma. From these 134 traumatic knee patients, 50% had small effusion, 26.9% had moderate effusion, and 4.5% had severe effusion. Clinically important effusion (moderate and severe effusion) detected by MRI was seen in 42 patients (31.3%). In 31 (73.8%) of the 42 patients with clinically important effusion a cruciate ligament lesion or meniscal tear was seen on MRI; a positive association (OR = 5.0; 95% CI, 2.2–11.1) of clinically important effusion with internal derangement of the knee could be shown (χ² 9.5, P=0.002).

History Taking and Physical Examination 

Of the 134 patients, for 128 patients (95.5%) both the history taking and the physical examination were available. Table 2 shows the association between items from history taking and physical examination with knee joint effusion.

Table 2. Number of Patients With Positive Test Result and Univariate Association of Items With Knee Joint Effusion

Variables	Patients Available n	Effusion Moderate/Severe⁎ (n=42)	Effusion Absent/Small⁎ (n=91)	OR	95% CI
History taking
Self-noticed swelling	128	32	40	4.8‡	2.0–11.6
Physical examination
Effusion fossa poplitea5	129	31	55	2.1†	0.9–5.0
Bakers cyst fossa poplitea5	129	5	7	1.7	0.5–5.6
Fluctuation/Minor effusion test5	127	6	7	2.0	0.6–6.4
Ballottement test5, 6	131	34	44	5.1‡	2.0–12.6

From history taking self-noticed swelling showed an association with effusion, with an OR of 4.8 (95% CI, 2.0–11.6) (see table 2). Two test results obtained by physical examination (effusion fossa poplitea and positive ballottement test) showed an association with effusion with an OR of 2.1 (95% CI, 0.9–5.0) and 5.1 (95% CI, 2.0–12.6), respectively.

Table 3 shows the multivariate association of items from history taking and physical examination with effusion.

Table 3. Multivariate Association (and 95% CI) of Items With Knee Joint Effusion

Variable From History Taking or Physical Examination	Effusion Moderate/Severe⁎ OR 95% CI
History taking
Self-noticed swelling	5.3†(2.1–13.6)
Physical examination
Ballottement test	5.7†(2.2–15.2)
Explained variance (%)	27.9

After multivariate modelling self-noticed swelling from history taking and the ballottement test from physical examination showed an independent association with the presence of effusion with an OR of 5.3 (95% CI, 2.1–13.6) and 5.7 (95% CI, 2.2–15.2), respectively.

Diagnostic Value of History Taking and Physical Examination 

In this study population the prevalence of effusion (prior probability) was 0.31 (42 of 134). Table 4 presents the diagnostic value of history taking and physical examination.

Table 4. Diagnostic Values (and 95% CI) of Isolated Symptoms and Signs, and Combinations of Symptoms and Signs, With Knee Joint Effusion (Prevalence=0.31/n=42)

Variable	n⁎	SE	SP	PVP	PVN	LR+	LR−
Isolated symptoms and signs
Self-noticed swelling	72	0.80	0.45	0.40	0.83	1.5	0.4†
		(0.68–0.92)	(0.35–0.39)	(0.29–0.51)	(0.73–0.94)	(1.1–1.9)	(0.2–0.9)
Ballottement test	78	0.83	0.49	0.43	0.86	1.6	0.3†
		(0.71–0.94)	(0.39–0.59)	(0.32–0.53)	(0.77–0.96)	(1.3–2.1)	(0.2–0.7)
Combination
Self-noticed swelling plus ballottement test	42	0.67	0.82	0.62	0.85	3.6†	0.4†
		(0.52–0.81)	(0.73–0.90)	(0.47–0.77)	(0.77–0.92)	(2.2–5.9)	(0.3–0.6)

Abbreviations: SE, sensitivity; SP, specificity.

The PVP of effusion increased from 0.31 to 0.40 (95% CI, 0.29–0.51) for self-noticed swelling and to 0.43 (95% CI, 0.32–0.53) for the ballottement test. The PVP increased to 0.62 (95% CI, 0.47–0.77) when self-noticed swelling from history taking and the ballottement test from physical examination were both present.

The probability of the absence of effusion increased from 0.69 to 0.83 (95% CI, 0.73–0.94) if self-noticed swelling was absent and to 0.86 (95% CI, 0.77–0.96) if the ballottement test was negative. Absence of self-noticed swelling together with a negative ballottement test did not further increase the PVN.

The isolated symptom self-noticed swelling and the ballottement test showed an LR+ of 1.5 (95% CI, 1.1–1.9) and 1.6 (95% CI, 1.3–2.1), respectively; altering the likelihood of effusion by only a small degree. The LR+ increased to a clinically important value of 3.6 (95% CI, 2.2–5.9) when the symptom self-noticed swelling from history taking and the ballottement test from physical examination were both present.

The symptom self-noticed swelling and the ballottement test had a clinically important LR− of 0.4 (95% CI, 0.2–0.9) and 0.3 (95% CI, 0.2–0.7), respectively. Absence of self-noticed swelling together with a negative ballottement test did not further decrease the LR−.

Back to Article Outline

Discussion 

The present study is the first to investigate the diagnostic value of history taking and physical examination for knee effusion seen on MRI, among patients in a primary care setting. Clinically important effusion was seen in 31.3% of these 134 patients.

In the present study the isolated symptom self-noticed swelling from history taking and the ballottement test from physical examination showed diagnostic value in detecting and moreover in excluding knee joint effusion. Combining the symptom from history taking and the sign from physical examination resulted in a more clinically important PVP and LR+. In this case GPs can predict with a 62% chance whether there is knee joint effusion, but still cannot be certain about the actual presence of effusion. Further, GPs can reasonably exclude (with an 83%–86% chance) knee joint effusion when self-noticed swelling and/or the ballottement test is negative, which could help to avoid unnecessary diagnostic interventions. However, Boks et al¹² showed that in case of absence of effusion some patients still have internal derangement of the knee.

To our knowledge, the only other study addressing the detection of knee joint effusion focused on the accuracy of sonographic examination. Wang et al¹⁵ reported a sensitivity, specificity, PVP and PVN of sonographic examination (compared with MRI) for detecting knee effusion. They concluded that sonographic examination can accurately detect effusion of the knee. Compared with Wang, our study showed equal sensitivity and specificity for self-noticed swelling and the ballottement test. Combining self-noticed swelling and the ballottement test in our study showed a lower sensitivity, but a higher specificity. Further, our study showed lower PVPs, but much higher PVNs for self-noticed swelling, the ballottement test, and the combination of both.

In the present study during history taking we asked the patient whether they had noticed a swelling of the knee; the answer to this question could be no, sometimes, or always. For the statistical analysis we dichotomized the answers into no versus sometimes and always. Alternatively, dichotomizing the answer into no and sometimes versus always did not improve the results.

Further, from physical examination only the ballottement test showed an association with knee joint effusion. Miller et al¹⁶ found a significant association between Baker cyst and effusion, whereas we did not. An explanation for this difference could be that we had only 12 patients with a Baker cyst and (in contrast to Miller et al¹⁶) we palpated for a Baker cyst instead of using MRI, whereby a small Baker cyst could be missed.

The literature⁵, ⁶ assumes that one can evaluate the presence of effusion with the minor effusion test and the ballottement test. However, our study shows that, of these 2 tests, only the ballottement test has an association with clinically important effusion. Although one might argue that the minor effusion test may be better in detecting minor effusion, we even found no association between the minor effusion test and small or minor effusion.

Discussion remains about interpretation of the ballottement test. Is the test positive when you hear a click⁵, ⁶ or is it positive when you feel a floating patella?⁵ In the present study we considered both a click and a floating patella to be a positive test result. In most of our patients a positive ballottement test concerned a floating patella (n=74) and only a few concerned a click (n=4).

For the analyses we dichotomized effusion seen on MRI into none to small and moderate to severe effusion, because Boks et al¹² reported that a small effusion was very common in asymptomatic knees and might not be clinically meaningful. Kolman et al³ also suggested that a small effusion (10mm or less in the lateral suprapatellar pouch seen on MRI) was not clinically meaningful, but rather a physiologic sign.

The period between trauma and MRI was 38 (range, 9–81) days. In this period, natural healing of lesions (at least of haemathrosis, but also of effusion) might have occurred, and some effusion might have changed from moderate or severe to small. Because history taking was assessed earlier than physical examination and MRI, the relationship between history taking and effusion on MRI could be underestimated.

Because we had a relatively small study population (N=134) we have chosen to use a cutoff point of 0.10 in the univariate analysis for inclusion in the multivariate model. However, the variables that remained in the multivariate model and were used for the diagnostic values both showed much lower P values (P<0.05). Still, the results presented here should be validated in a larger study population in general practice.

The Dutch Clinical Guideline Traumatic knee disorders¹⁷ for GPs reported that in case of effusion within a few hours after the trauma, in the presence of a positive ballottement test or locking of the knee there is suspicion of internal derangement of the knee; they recommend a wait and see policy unless there is locking of the knee or suspicion of a fracture. Some data are available concerning the clinical value of effusion. Duncan et al² describe a positive correlation between effusion and internal derangement of the knee. In addition, Kolman³ reported that symptomatic patients with no significant effusion in the lateral aspect of the suprapatellar pouch on MRI are free of internal derangement in 86% of the cases. Our study confirmed also a positive association between clinically important effusion and internal derangement of the knee. Additionally, and more important for GPs, is the question whether there is also an association between history taking/physical examination (focused on effusion) and internal derangement of the knee. We found a small relationship between positive self-noticed swelling together with a positive ballottement test and internal derangement of the knee (univariate logistic regression, OR=1.9; P=0.094), predicting a 59% chance of internal derangement of the knee (data not shown). The presence of only the symptom self-noticed swelling showed almost the same relationship with internal derangement of the knee, but the presence of only a positive ballottement test did not show this relationship (data not shown). Based on the present results we recommend the following: if GPs suspect effusion of the knee, because the patient has noticed swelling of the knee and the ballottement test is positive, they should consider the possibility of internal derangement of the knee. A wait and see policy is recommended unless there is locking of the knee or suspicion of a fracture. However, because of the association between effusion and internal derangement of the knee, GPs might use the predicted chance of effusion as a selection criterion for MRI and/or referral to secondary care.

Back to Article Outline

Conclusions 

Our study indicates that GPs can reasonably exclude knee joint effusion in patients with a knee trauma based on the absence of self-noticed swelling and/or a negative ballottement test. This could argue against further diagnostic interventions, but the GP has to bear in mind that some patients with knee joint effusion could be missed. The data also show that GPs can predict knee joint effusion with a maximum of only 62% certainty, based on the presence of self-noticed swelling together with a positive ballottement test. In addition, our study results support that knee joint effusion in primary care patients is associated with internal derangement of the knee.

Suppliers

Back to Article Outline

Appendix 

Appendix 1. Definition of Questions From History Taking and Tests From Physical Examination

History Taking
Question	Answer
Present symptoms
Do you have swelling of the knee?	Always; sometimes; no.

Back to Article Outline

References 

1. van der Linden M, Westert G, de Bakker D, Schellevis F. [Second National Study into diseases and actions in general practice: complaints and disorders in the population and in general practice] [Dutch]. Utrecht/Bilthoven: NIVEL/RIVM 2004. Available at: www.nivel.nl/pdf/ns2_r1_h02.pdf. Accessed November 20, 2008.
   • View In Article
2. Duncan JB, Hunter R, Purnell M, Freeman J. Injured stable knee with acute effusion: MRI evaluation. J South Orthop Assoc. 1996;5:13–19
   • View In Article
   • MEDLINE
3. Kolman BH, Daffner RH, Sciulli RL, Soehnlen MW. Correlation of joint fluid and internal derangement on knee MRI. Skeletal Radiol. 2004;33:91–95
   • View In Article
   • MEDLINE
   • CrossRef
4. Cardol M, van Dijk L, de Jong J, de Bakker D, Westert G. [Second National Study into diseases and actions in general practice. General practitioner care: what does the gate keeper?] [Dutch] Utrecht/Bilthoven: NIVEL/RIVM 2004. Available at: www.nivel.nl/pdf/ns2_r2_h00.pdf. Accessed November 20, 2008.
   • View In Article
5. Hoppenfield S. Physical examination of the spine and extremities. East Norwalk: Appleton Century Crofts; 1976;
   • View In Article
6. Reider B. The orthopaedic physical examination. In: Philadelphia: WB Saunders; 1999;p. 202–245
   • View In Article
7. Noyes FR, Cummings JF, Grood ES, Walz-Hasselfeld KA, Wroble RR. The diagnosis of knee motion limits, subluxations, and ligament injury. Am J Sports Med. 1991;19:163–171
   • View In Article
   • MEDLINE
   • CrossRef
8. Oberlander MA, Shalvoy RM, Hughston JC. The accuracy of the clinical knee examination documented by arthroscopy (A prospective study). Am J Sports Med. 1993;21:773–778
   • View In Article
   • MEDLINE
   • CrossRef
9. Schweitzer ME, Falk A, Berthoty D, Mitchell M, Resnick D. Knee effusion: normal distribution of fluid. AJR Am J Roentgenol. 1992;159:361–363
   • View In Article
   • MEDLINE
10. Heintjes EM, Berger MY, Koes BW, Bierma-Zeinstra SM. Knee disorders in primary care: design and patient selection of the HONEUR knee cohort. BMC Musculoskelet Disord. 2005;6:45
    • View In Article
    • MEDLINE
    • CrossRef
11. Tegner Y, Lysholm J. Rating systems in the evaluation of knee ligament injuries. Clin Orthop Relat Res. 1985;198:43–49
    • View In Article
12. Boks SS, Vroegindeweij D, Koes BW, Hunink MM, Bierma-Zeinstra SM. Magnetic resonance imaging abnormalities in symptomatic and contralateral knees: prevalence and associations with traumatic history in general practice. Am J Sports Med. 2006;34:1984–1991
    • View In Article
    • MEDLINE
    • CrossRef
13. Hunink MG. Decision making in the face of uncertainty and resource constraints: examples from trauma imaging. Radiology. 2005;235:375–383
    • View In Article
    • MEDLINE
    • CrossRef
14. Gallagher EJ. Clinical utility of likelihood ratios. Ann Emerg Med. 1998;31:391–397
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (48 KB)
    • CrossRef
15. Wang CY, Wang HK, Hsu CY, Shieh JY, Wang TG, Jiang CC. Role of sonographic examination in traumatic knee internal derangement. Arch Phys Med Rehabil. 2007;88:984–987
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (161 KB)
    • CrossRef
16. Miller TT, Staron RB, Koenigsberg T, Levin TL, Feldman F. MR imaging of Baker cysts: association with internal derangement, effusion, and degenerative arthropathy. Radiology. 1996;201:247–250
    • View In Article
    • MEDLINE
17. Plas CG van der HA. [Guideline regarding traumatic knee problems] [Dutch]. Huisarts en wetenschap. 1998;41:
    • View In Article

• a GE Signa Horizon LX 1.0 Tesla scanner; GE Medical Systems, 5486 N Lake Dr, Milwaukee, WI 53217.
• b Logistic regression analysis, version 11.0; SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.