Inflammatory low back pain: High negative predictive value of contrast-enhanced color Doppler ultrasound in the detection of inflamed sacroiliac joints




To determine the value of microbubble contrast agents for color Doppler ultrasound (CDUS) compared with magnetic resonance imaging (MRI) in the detection of active sacroiliitis.


An observational case-control study of 103 consecutive patients (206 sacroiliac [SI] joints) with inflammatory low back pain according to the Calin criteria and 30 controls (60 SI joints) without low back pain was conducted at the University Hospital of Innsbruck. All patients and controls underwent unenhanced and contrast-enhanced CDUS and MRI of the SI joints. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of unenhanced and contrast-enhanced CDUS were evaluated.


Forty-three patients (41%) with 70 of 206 SI joints (34%) and none of the controls nor the 60 control SI joints demonstrated active sacroiliitis on MRI. Unenhanced CDUS showed a sensitivity of 17%, a specificity of 96%, a PPV of 65%, and an NPV of 72%; contrast-enhanced CDUS showed a sensitivity of 94%, a specificity of 86%, a PPV of 78%, and an NPV of 97%. Detection of vascularity in the SI joint was increased by contrast administration (P < 0.0001). Clustered receiver operating curve analysis demonstrated that enhanced CDUS (Az = 0.89) was significantly better than unenhanced CDUS (Az = 0.61) for the diagnosis of active sacroiliitis verified by MRI (P < 0.0001; 2-sided test).


Microbubble contrast-enhanced CDUS is a sensitive technique with a high NPV for detection of active sacroiliitis compared with MRI.


Sacroiliitis is a frequent and early manifestation of the spondylarthropathies, including ankylosing spondylitis, psoriatic arthritis, reactive arthritis, enteropathic spondylarthropathy, and undifferentiated spondylarthropathy. As a group, the prevalence of these inflammatory spondylarthropathies is as high as 0.5–1.9% (1). Early diagnosis and treatment of sacroiliitis can improve clinical outcome. Clinical diagnosis and physical examination, however, are not very specific.

Jarvik and Deyo reported that clinical tests of sacroiliac (SI) joint tenderness are poorly reproducible and inaccurate in distinguishing sacroiliitis from mechanical spinal conditions (2). Serum markers, such as C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), are not necessarily elevated (3). Although not crucial for the diagnosis of spondylarthritis according to the criteria of the European Spondylarthropathy Study Group, magnetic resonance imaging (MRI) of sacroiliitis may help, especially early in the course of the disease. MRI can demonstrate early predestructive alterations of sacroiliitis, and thus provide an early diagnosis of sacroiliitis (4, 5). The availability of MRI, however, is limited. The technique is time consuming and costly (4, 6, 7), and may not be applied to all patients with inflammatory low back pain and suspected sacroiliitis in clinical practice.

Color Doppler ultrasound (CDUS) is a technology widely used for detection of blood flow, especially of slow flow and flow in such small vessels as neovessels. Power Doppler ultrasound (PDUS) is theoretically more sensitive to flow and independent of the angle of insonation, but it is also more sensitive to artifacts compared with CDUS. The detection of low-volume blood flow is even improved when CDUS/PDUS is enhanced with a microbubble contrast agent. Microbubble contrast agents for CDUS have recently been shown to improve the detection of increased vascularity in inflammatory arthritic diseases (8–10).

In this study we determined the values of unenhanced and contrast-enhanced CDUS in the detection of active sacroilitiis compared with contrast-enhanced MRI.



One hundred three consecutive patients (80 women, 23 men; mean ± SD age 47.5 ± 31.1 years, range 26–65 years; 206 SI joints) who fulfilled at least 4 criteria for inflammatory low back pain according to Calin et al (11) underwent MRI and sonographic evaluation after having given informed and written consent according to the declaration of Helsinki. Out of the criteria established by Calin et al, all patients had a slow onset of disease, symptoms persisting >3 months, associated morning stiffness, and improvement with exercise. In addition, 34 of the 103 patients (33%) were younger than 40 years at the time of presentation. Of these, 74.8% were finally diagnosed with spondylarthropathy (17 with ankylosing spondylitis, 5 with psoriatic arthritis, 1 with reactive arthritis, and 54 with undifferentiated spondylarthritis). In the group, ESR was 20.45 ± 19.0 mm/hour (range 2–90; normal defined as <15 mm/hour), CRP was 0.99 ± 0.57 mg/dl (range 0.70–3.54; normal defined as <0.07 mg/dl), and 28 of the 103 patients with inflammatory low back pain (27.1%) were HLA–B27 positive.

Thirty volunteers (20 women, 10 men; mean ± SD age 26.3 ± 5.9 years, range 20–35 years; 60 SI joints) without suspicion of inflammatory low back pain and who had an MRI available because of an independent diagnosis served as a control group.

Magnetic resonance imaging.

MRI was performed with a Magnetom Vision 1.5 Tesla unit (Siemens, Erlangen, Germany) using the following pulse sequences: T1-weighted semiaxial spin-echo (SE), T2-weighted turbo spin-echo, T1-weighted SE with fat saturation, T1-weighted semicoronal SE, turbo inversion recovery magnitude (TIRM), and semicoronal and semiaxial T1 fat saturation after administration of gadolinium contrast agent (Magnevist, Schering, Germany). MRI images were evaluated by an experienced radiologist who was unaware of the sonographic results.

Sonographic evaluation.

Sonographic evaluation was performed with the patient in the prone position. Gray-scale ultrasound was performed first to identify the bony spinous processes in the midline and the posterior part of the SI joints (12). CDUS imaging was performed with a Sequoia 512 unit (Acuson, Mountain View, CA) fitted with a 6C2 probe, operating at a Doppler frequency of 2.5–5 MHz. We used standardized machine settings for the CDUS examination as follows: color box restricted to the area of the SI joints; velocity scale of 3.2 cm/second; color Doppler gain of 52 dB; one focal zone. CDUS examination was performed to detect vascularization, which was defined as color-flow signals in the area of the SI joints.

Once the midline was identified and the iliac bones were depicted, the transducer was moved caudal to the level of the first sacral foramen where the hypoechoic cleft of the examined SI joints could be delineated. CDUS examination was performed to detect vascularization, which was defined as color-flow signals in the area of the SI joints. For each SI joint, the area with the highest number of CDUS flow signals was counted for statistical analysis. We did not use the mean calculation of color flow signals, but the scan with the highest number of CDUS flow signals.

The ultrasound contrast agent, SonoVue (Bracco, Milano, Italy), is administered intravenously. SonoVue is an aqueous suspension of stabilized sulfur hexafluoride microbubbles. The bubble size is between 1 and 10 μm, allowing the agent to pass in the circulation through the pulmonary capillary bed and serve as a strong reflector for ultrasound. Contraindications for the application of SonoVue are known hypersensitivity to sulfur hexafluoride or to any components of SonoVue, right-to-left shunt, severe pulmonary hypertension (pulmonary artery pressure > 90 mm Hg), uncontrolled systemic hypertension or adult respiratory distress syndrome, chronic obstructive pulmonary disease, severe congestive heart failure, severe arrhythmia, or pregnancy.

The ultrasound contrast agent was administered as an intravenous bolus to a maximum dose of 4.8 ml and flushed with 5–10 ml saline. After contrast administration, CDUS of the SI joints was performed with a slow sweep from the lumbosacral region (spinous process of L5–S1) to the caudal portion of the SI joints. Images were digitally stored. For each SI joint, the area with the highest number of CDUS flow signals was counted for statistical analysis. CDUS findings of vascularity of the SI joint were rated on a subjective 4-point scale for both unenhanced and contrast-enhanced CDUS as used previously for grading of finger joints in patients with rheumatoid arthritis (Table 1) (10). A spectral Doppler tracing was obtained to confirm that each color Doppler signal represented true arterial or venous flow.

Table 1. Color Doppler ultrasound: Subjective grading scale of vascularity in the sacroiliac joint (a 4-point scale)
Number of flow signalsGrade of vascularityRating

The person performing the ultrasound was blinded to the clinical data of the patient and the MRI findings and was not aware of the history or clinical examination of the patients.

Statistical analysis.

The rating scores obtained from unenhanced and contrast-enhanced CDUS were evaluated with MRI as the standard. The degree of vascularization observed in the SI joints of controls was compared with that observed in patients with sacroiliitis using the Mann-Whitney 2-sample statistics. Differences between observed vascularity of unenhanced and contrast-enhanced CDUS were evaluated with Wilcoxon's matched-pairs signed-rank test. To evaluate the diagnostic accuracy of CDUS in our patient population, and to compensate for the lack of statistical independence between the 2 SI joints in each patient, the subjective sonographic ratings of unenhanced and contrast-enhanced CDUS were compared with a clustered receiver operating characteristic (ROC) analysis (13).


Forty-three of 103 patients with inflammatory low back pain (42%) (70 of 206 joints [34%]) demonstrated active sacroiliitis on MRI. Unenhanced CDUS detected 8 of these 43 patients (19%; 11 of 70 joints [16%]) and contrast-enhanced CDUS detected 41 of the 43 patients (95%; 66 of 70 joints [94%]) (Figure 1). Thus, detection of hypervascularity in the MRI-defined inflamed SI joints was improved after contrast administration (P < 0.001). Clustered ROC analysis of 206 SI joints in 103 patients confirmed that the enhanced ultrasound (Az = 0.89) was better than unenhanced ultrasound (Az = 0.61) for the detection of active sacroiliitis as defined by MRI (P < 0.0001; 2-sided test) (Figure 2).

Figure 1.

A 45-year-old female patient with magnetic resonance-verified bilateral active sacroiliitis. A, Unenhanced color Doppler ultrasound (CDUS) demonstrates no color flow signals in the sacroiliac joints (arrows). B, Contrast-enhanced CDUS demonstrates grade 2 vascularity on the left side (arrows) and grade 3 vascularity on the right side (arrows), suspicious for bilateral sacroiliitis. SP = spinous process.

Figure 2.

Receiver operating characteristic (ROC) curves for unenhanced and contrast-enhanced color Doppler ultrasound (CDUS) in the detection of active sacroiliitis in 206 sacroiliac joints of 103 patients with magnetic resonance-verified active sacroiliitis. ROC points on the curve for unenhanced CDUS are marked with solid circles and ROC points on the curve for contrast-enhanced CDUS are marked with open circles. The estimated area, Az, for unenhanced CDUS is 0.61 and for contrast-enhanced CDUS is 0.89 (P < 0.0001; 2-sided test).

None of the controls had MRI-defined inflamed SI joints, and neither unenhanced nor enhanced CDUS detected hypervascularity in the control SI joints. Controls without low back pain demonstrated only grade 0 and 1 vascularity. Vascularity was significantly lower in the SI joints of the controls as compared with those of patients with sacroiliitis (P < 0.0001).

For detection of active sacroiliitis as found by MRI, unenhanced CDUS showed a sensitivity of 17%, a specificity of 96%, a positive predictive value (PPV) of 65%, and a negative predictive value (NPV) of 72% and contrast-enhanced CDUS demonstrated a sensitivity of 94%, a specificity of 86%, a PPV of 78%, and an NPV of 97%.

No side effects were noted after administration of the ultrasound contrast agent. The mean ± SD examination time for CDUS was 22 ± 12.5 minutes (range 17.5–33.4 minutes).


Our results demonstrate that the visualization of hypervascularization in the SI joints can be improved by microbubble contrast administration compared with unenhanced CDUS. Contrast-enhanced CDUS is a safe and highly sensitive technique for the detection of active sacroiliitis.

A previous study by Arslan et al studying unenhanced CDUS in a group of 21 patients with active sacroiliitis had already demonstrated vascularity around or inside the SI joint in 10 patients (48%), but also in some of the patients with osteoarthritis and in the controls (12). In our hands, unenhanced CDUS detected hypervascularity in the SI joints of 8 of 43 patients (19%) with active sacroiliitis. The difference in the unenhanced CDUS findings is likely explained by the fact that we included only those color flow signals that were clearly obtained in the SI joint (the hypoechoic cleft seen on gray-scale ultrasound).

Several studies have shown that the use of a contrast agent improves detection of blood flow in the inflamed synovium (9, 10). From studies in ankylosing spondylitis, we know that an increased vascularity in the SI joint and increased levels of proangiogenic factors can be detected in the sera of these patients, and may even be associated with increased disease activity (14–16). The use of contrast agents has increased the sensitivity of CDUS to detect even slow flow and flow in small vessels. Indeed, using CDUS, we found an increased rate of hypervascularity in the SI joints of patients with active sacroiliitis as detected by MRI, but not in controls or patients with low back pain but normal MRI findings. However, false-positive results were obtained and we cannot explain them. Scoring of MRI positivity concerning sacroiliitis has improved recently by new scoring systems and by the use of Gadolinium. Indeed, 3 of the 4 joints that were MRI positive and CDUS negative showed signs of inflammation limited to the anterosuperior portion of the sacroiliac joint area. Another one showed inflammation at the inferior portion of the sacroiliac joint with joint space narrowing and spur formation posteriorly. Therefore, inflammation at the anterosuperior portion only and prominent spur formation dorsally can be considered limitations for the CDUS examination, because of limited ultrasound beam penetration. For these cases with enhanced CDUS findings but negative MRI, pathohistologic examinations and a long-term followup will be necessary to clarify the CDUS findings.

The advantage of using MRI in patients who meet the Calin criteria is that the lumbar spine can also be examined for spondylitis, but MRI is expensive and in our country it is not available as a screening tool for all patients with inflammatory low back pain. MRI is also limited in patients with metal implants, pacemakers, or claustrophobia. Nevertheless, MRI is superior to plain radiography and computed tomography, because it can detect active sacroiliitis at an early phase of the disease (17, 18). As opposed to these modalities, contrast-enhanced CDUS is a relatively simple, portable, and less-expensive imaging modality that could be used as a bedside diagnostic tool (19).

There are several limitations to our study. First, the CDUS examinations were performed by a single investigator. Second, the grading of vascularity was subjective. Although a study by Walther et al found good correlation between subjective grading of vascularity in the pannus of the knee and digital image evaluation (20), objective assessment by computer-assisted quantification may further improve the value of contrast-enhanced CDUS in the detection of sacroiliitis. Third, pathohistologic correlation is still considered to be the true standard for defining active sacroiliitis, at least for scientific evaluations of new diagnostic techniques (18).

Because of the high negative predictive value of contrast-enhanced CDUS in the detection of active sacroiliitis, it is unlikely that MRI will show signs of active sacroiliitis when contrast-enhanced CDUS is normal. The application of contrast-enhanced CDUS may prevent the performance of MRI studies in a number of patients with typical history and clinical signs of inflammatory low back pain. In our country, the cost of contrast-enhanced CDUS is estimated to be less than one-third of that for contrast-enhanced MRI studies. Thus, expensive MRI could be preserved for patients with positive contrast-enhanced CDUS suspicious of active sacroiliitis.

In summary, contrast-enhanced CDUS is a promising and readily available imaging modality with a high negative predictive value for active sacroiliitis.