- Email: [email protected]
Cerebrospinal fluid angiotensin-converting enzyme for diagnosis of neurosarcoidosis
Journal of Neuroimmunology 285 (2015) 1–3
Contents lists available at ScienceDirect
Journal of Neuroimmunology journal homepage: www.elsevier.com/locate/jneuroim
Short communication
Cerebrospinal fluid angiotensin-converting enzyme for diagnosis of neurosarcoidosis Claire Bridel a,⁎, Delphine S. Courvoisier b, Nicolas Vuilleumier c, Patrice H. Lalive a,c a b c
Department of Clinical Neurosciences, Division of Neurology, Unit of Neuroimmunology and Multiple Sclerosis, University Hospital of Geneva, Switzerland Quality of Care Unit, University Hospital of Geneva, Switzerland Department of Genetics and Laboratory Medicine, Laboratory Medicine Service, University Hospital of Geneva, Switzerland
a r t i c l e
i n f o
Article history: Received 22 April 2015 Received in revised form 14 May 2015 Accepted 18 May 2015 Keywords: Neurosarcoidosis Cerebrospinal fluid Angiotensin-converting enzyme Granulomatous diseases Inflammatory diseases of the central nervous system
a b s t r a c t Background: Neurosarcoidosis (NS) is a rare condition that may mimic central nervous system (CNS) infection, neoplasia and other inflammatory disorders of the CNS such as multiple sclerosis, encephalitis and vasculitis. Diagnosis is challenging in cases with minimal or absent systemic involvement. Cerebrospinal fluid (CSF) angiotensin-converting enzyme (c-ACE) has been claimed as a valuable diagnostic tool for NS. However, there is little data evaluating its performance in routine clinical practice. Findings: We performed a monocentric, retrospective, chart-based study including all patients investigated with a lumbar puncture and c-ACE dosage for suspected NS between 01/01/2006 and 31/12/2012 at the Geneva University Hospital. Receiver-operating characteristic (ROC) curve and area under the curve (AUC) were performed to calculate the optimal cut-off value of c-ACE and to determine the discriminative ability of c-ACE. Of the 440 patients included in the study, 9 were diagnosed with NS on the basis of tissue biopsy. Mean c-ACE was not significantly different between NS and non-NS patients. With a cut-off value of 2 (0–2 vs ≥3), sensitivity and specificity of c-ACE were 66.7% and 67.3%, respectively. Conclusions: In our clinical setting, the sensitivity and specificity of c-ACE for NS diagnosis were relatively poor and of little clinical utility. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Sarcoidosis is a rare chronic granulomatous systemic disorder affecting multiple organs including the nervous system (Chen and Moller, 2011). Pulmonary involvement occurs in 90% of patients, and diagnosis can be established by transbronchial or mediastinal lymph node biopsy (Israel-Biet and Valeyre, 2013). Body fluid biomarkers supporting diagnosis of systemic sarcoidosis (SS) include elevated serum angiotensinconverting enzyme (s-ACE), increased (N3.5) CD4 +/CD8 + ratio in broncho-alveolar lavage, and elevated 24 h-calciuria. Central nervous system (CNS) involvement (neurosarcoidosis, NS) occurs in 10–25% of patients (Chen and Moller, 2011) and is isolated in 1% (Vargas and Stern, 2010). NS may mimic CNS infection, neoplasia and other
Abbreviations: CNS, central nervous system; NS, neurosarcoidosis; SS, systemic sarcoidosis; ACE, angiotensin-converting enzyme; CSF, cerebrospinal fluid; c-ACE, cerebrospinalfluid angiotensin-converting enzyme; s-ACE, serum angiotensin-converting enzyme; ROC, receiver-operating characteristic; AUC, area under the curve; WBC, white blood cells. ⁎ Corresponding author at: Department of Clinical Neurosciences, Division of Neurology, University Hospital of Geneva, Rue Gabrielle-Perret-Gentil 4, 12011 Geneva, Switzerland. E-mail addresses: [email protected] (C. Bridel), [email protected] (D.S. Courvoisier), [email protected] (N. Vuilleumier), [email protected] (P.H. Lalive).
http://dx.doi.org/10.1016/j.jneuroim.2015.05.020 0165-5728/© 2015 Elsevier B.V. All rights reserved.
inflammatory disorders of the CNS such as multiple sclerosis, encephalitis and vasculitis (Reske et al., 2005). Diagnosis is challenging in cases with minimal or absent systemic involvement. Definite diagnosis requires CNS biopsy, which is not always achievable. Probable NS is defined as suggestive CNS lesions together with evidence of SS (positive systemic tissue biopsy and/or positive chest imaging together with elevated serum angiotensin converting enzyme (s-ACE)) (Zajicek et al., 1999; Marangoni et al., 2006). Cerebrospinal fluid (CSF) ACE (c-ACE) has been claimed as a valuable diagnostic tool for NS (Oksanen et al., 1985; Tahmoush et al., 2002; Reske et al., 2005). However, there is little data evaluating its performance in routine clinical practice, and results of studies are conflicting (Dale and O'Brien, 1999; Tahmoush et al., 2002). No specific biomarker is available to support NS diagnosis, underscoring the importance of c-ACE validation. In many hospitals, c-ACE is performed routinely when NS is suspected. We questioned its pertinence as a biomarker of NS, and performed a retrospective study to evaluate its clinical usefulness.
2. Materials and methods 2.1. Study design This is a retrospective monocentric chart-based study.
2
C. Bridel et al. / Journal of Neuroimmunology 285 (2015) 1–3
c-ACE levels were not significantly different in the NS group compared to the non-NS group (Table 1A). The ROC-curve (Fig. 1A) showed low diagnosis accuracy, with an AUC of 0.65. The most discriminative cut-off point was a c-ACE of 2, yielding a sensitivity of 66.7% (95% CI 0.30–0.93) and a specificity of 67.3% (95% CI 0.63–0.72) (Fig. 1B). However, no cut-off provided a good likelihood ratio. Note that given the rarity of NS in our sample (9 patients), the estimate of sensitivity was imprecise, as reflected in the large 95% CI. s-ACE levels were significantly different between groups, with a higher mean value in the NS group. However, while all 9 NS patients had values, 158 non-NS patients had missing values.
2.2. Patients Consecutive patients investigated at the Geneva University Hospital for clinically suspected NS between January 2006 and December 2012 were included. Clinical, demographical, radiological and biological information (including leucorachia, proteinorachia, intrathecal immunoglobulin synthesis, c-ACE, s-ACE), final diagnosis and, when pertinent, patho-histology, were collected. NS was considered present if (1) the patient had histologically proven NS upon CNS biopsy (definite NS), or (2) the patient had suggestive CNS lesions together with evidence of SS (probable NS). This study received approval from the local Ethics Committee.
4. Discussion
2.3. c-ACE activity measure
In this study, we find that mean c-ACE is not significantly higher in the NS group compared to the non-NS group, in contrast to what was previously reported in an American case–control study including 11 patients with probable NS and 207 controls (Tahmoush et al., 2002). Tahmoush et al. selected controls consisting of patients with other CNS and peripheral nervous system (PNS) diseases (Tahmoush et al., 2002), whereas our control group consisted of any patient with suspected NS, which is closer to a clinically relevant situation. We find that c-ACE is of poor clinical interest for NS diagnosis. Indeed, even the best cut-off value (0–2 vs. ≥ 3) only provided a sensitivity of 66.7%% with a specificity of 67.3%. This finding is actually slightly better than previously published results. Tahmoush et al. reported a sensitivity of 55% (Tahmoush et al., 2002) and Dale and O'Brien a sensitivity of 24% (Dale and O'Brien, 1999). Our study has two main limitations. First, its retrospective nature may have led to some information bias, with NS patients being investigated more closely, as exemplified by the missing data for s-ACE in non-NS patients. However, given the rarity of this condition in Western countries, a prospective study is difficult to achieve. Second, the degree of NS suspicion varies greatly among the patients included in this study.
The measure of ACE activity in CSF was assessed with the colorimetric ACE kinetic test reagents from Bühlmann Laboratories (AG, Switzerland) using the RX Daytona™ chemistry analyzer (Randox, Minnesota, USA), according to the manufacturers' instructions. Results are expressed in units per liter (U/L). Inter-series variability was 3.5% (n = 15) and the limit of quantification was 0.3 U/L. 2.4. Statistical analysis Kruskal–Wallis test was used to compare continuous variables, and Fisher exact test was used to compare categorical variables. Receiveroperating characteristic (ROC) curve and area under the curve (AUC) were performed to choose the optimal cut-off value of c-ACE and to determine the discriminative ability (sensitivity and specificity) of c-ACE as a diagnostic marker for NS. Gold standard for NS diagnosis was characteristic patho-histology of tissue biopsy. Statistical analysis was performed using STATA v13.1 (StataCorp., College Station, TX). 3. Results Of the 440 patients who underwent a lumbar puncture with c-ACE measure for suspected NS between 2006 and 2012 at the Geneva University Hospital (Table 1A), 8 had probable NS with histologicallyproven SS and 1 had definite isolated NS. Patients were close to the same age, and there were significantly more males in the NS group (Table 1A). A complete description of the 9 NS cases is found in Table 1B.
5. Conclusion In this retrospective study of 440 patients with 9 NS cases, we found that c-ACE had a relatively low sensitivity and specificity. This study does not support using c-ACE as a diagnostic laboratory test.
Table 1 Sample characteristics. A
N Age, mean (SD) Gender, N (% female) c-ACE, mean (SD) s-ACE, mean (SD)
NS
Other diagnosis
P
9 43.0 (13.4) 2 (22.2%) 4.3 (4.7) 77.2 (49.5)
431 48.5 (16.7) 261 (60.6%) 2.3 (2.7) 32.8 (18.1)
0.31 0.03 0.11 0.001
B Case
1
2
3
4
5
6
7
8
9
Age at diagnosis Gender s-ACE c-ACE CSF WBC (/mm3) CSF protein (N0.45 g/L) Intrathecal IgG synthesis Systemic sarcoidosis
56 M 9 1 6 Yes No No
48 M 86 0 3 Yes No Yes
28 M 29 3 2 No No Yes
29 F 124 3 5 Yes nd Yes
30 M 58 12 2 Yes No Yes
26 F 168 6 5 Yes nd Yes
51 M 61 4 5 Yes No Yes
68 M 105 7 nd nd nd Yes
47 M 49 1 4 No nd Yes
A. Clinical and biological characteristics of the patients included in the study. c-ACE, cerebrospinal fluid angiotensin-converting enzyme; p-ACE, plasma angiotensin-converting enzyme. B. Clinical and biological characteristics of neurosarcoidosis patients. c-ACE, cerebrospinal fluid angiotensin-converting enzyme; s-ACE, serum angiotensin-converting enzyme; CSF WBC, cerebrospinal fluid white blood cells; CSF protein, cerebrospinal fluid protein; IgG, immunoglobulin G.
C. Bridel et al. / Journal of Neuroimmunology 285 (2015) 1–3
3
A
B Cut-off value of c-ACE
Sensitivity (95% CI)
Specificity (95% CI)
Likelihood ratio
0
0.78 (0.39-0.97) 7/9 patients
0.34 (0.30-0.39) 148/431
1.18
1
0.66 (0.30-0.93) 6/9 patients
0.51 (0.46-0.56) 219/431 patients
1.35
2
0.66 (0.30-0.93) 6/9 patients
0.67 (0.63-0.72) 290/431 patients
2.00
3
0.44 (0.14-0.79) 4/9 patients
0.77 (0.73-0.81) 331/431 patients
1.91
4
0.33 (0.07-0.70) 3/9 patients
0.84 (0.80-0.87) 360/431 patients
2.07
Fig. 1. Discriminative ability of c-ACE. A. Receiver-operating characteristic (ROC) curve and area under ROC curve (AUC) of c-ACE. The ROC-curve shows low diagnosis accuracy of c-ACE, with an AUC of 0.65. B. Cut-off values of c-ACE and their associated sensitivity, specificity and positive likelihood ratio. Sensitivity, specificity and likelihood-ratio of c-ACE, according to the cut-off value of c-ACE. c-ACE, cerebrospinal fluid angiotensin-converting enzyme.
Author's contribution CB designed the study, collected the data, analyzed the data, performed the statistical analysis and wrote the manuscript. DSC checked the statistical analysis and contributed to the manuscript preparation. NV supervised the c-ACE analysis and revised the manuscript. PHL designed the study and revised the manuscript.
Competing interests Dr C. Bridel, Dr D.S. Courvoisier, DR N. Vuilleumier and Dr P.H. Lalive report no competing interests.
Acknowledgment We thank Lorraine Gibert and Manon Wehrli for their help in collecting data.
References Chen, E.S., Moller, D.R., 2011. Sarcoidosis—scientific progress and clinical challenges. Nat. Rev. Rheumatol. 7 (8), 457–467. Dale, J.C., O'Brien, J.F., 1999. Determination of angiotensin-converting enzyme levels in cerebrospinal fluid is not a useful test for the diagnosis of neurosarcoidosis. Mayo Clin. Proc. 74 (5), 535. Israel-Biet, D., Valeyre, D., 2013. Diagnosis of pulmonary sarcoidosis. Curr. Opin. Pulm. Med. 19 (5), 510–515. Marangoni, S., Argentiero, V., Tavolato, B., 2006. Neurosarcoidosis. Clinical description of 7 cases with a proposal for a new diagnostic strategy. J. Neurol. 253 (4), 488–495. Oksanen, V., et al., 1985. Angiotensin converting enzyme in cerebrospinal fluid: a new assay. Neurology 35 (8), 1220–1223. Reske, D., Petereit, H.F., Heiss, W.D., 2005. Difficulties in the differentiation of chronic inflammatory diseases of the central nervous system—value of cerebrospinal fluid analysis and immunological abnormalities in the diagnosis. Acta Neurol. Scand. 112 (4), 207–213. Tahmoush, A.J., et al., 2002. CSF-ACE activity in probable CNS neurosarcoidosis. Sarcoidosis Vasc. Diffuse Lung Dis. 19 (3), 191–197. Vargas, D.L., Stern, B.J., 2010. Neurosarcoidosis: diagnosis and management. Semin. Respir. Crit. Care Med. 31 (4), 419–427. Zajicek, J.P., et al., 1999. Central nervous system sarcoidosis—diagnosis and management. Q JM 92 (2), 103–117.
Contents lists available at ScienceDirect
Journal of Neuroimmunology journal homepage: www.elsevier.com/locate/jneuroim
Short communication
Cerebrospinal fluid angiotensin-converting enzyme for diagnosis of neurosarcoidosis Claire Bridel a,⁎, Delphine S. Courvoisier b, Nicolas Vuilleumier c, Patrice H. Lalive a,c a b c
Department of Clinical Neurosciences, Division of Neurology, Unit of Neuroimmunology and Multiple Sclerosis, University Hospital of Geneva, Switzerland Quality of Care Unit, University Hospital of Geneva, Switzerland Department of Genetics and Laboratory Medicine, Laboratory Medicine Service, University Hospital of Geneva, Switzerland
a r t i c l e
i n f o
Article history: Received 22 April 2015 Received in revised form 14 May 2015 Accepted 18 May 2015 Keywords: Neurosarcoidosis Cerebrospinal fluid Angiotensin-converting enzyme Granulomatous diseases Inflammatory diseases of the central nervous system
a b s t r a c t Background: Neurosarcoidosis (NS) is a rare condition that may mimic central nervous system (CNS) infection, neoplasia and other inflammatory disorders of the CNS such as multiple sclerosis, encephalitis and vasculitis. Diagnosis is challenging in cases with minimal or absent systemic involvement. Cerebrospinal fluid (CSF) angiotensin-converting enzyme (c-ACE) has been claimed as a valuable diagnostic tool for NS. However, there is little data evaluating its performance in routine clinical practice. Findings: We performed a monocentric, retrospective, chart-based study including all patients investigated with a lumbar puncture and c-ACE dosage for suspected NS between 01/01/2006 and 31/12/2012 at the Geneva University Hospital. Receiver-operating characteristic (ROC) curve and area under the curve (AUC) were performed to calculate the optimal cut-off value of c-ACE and to determine the discriminative ability of c-ACE. Of the 440 patients included in the study, 9 were diagnosed with NS on the basis of tissue biopsy. Mean c-ACE was not significantly different between NS and non-NS patients. With a cut-off value of 2 (0–2 vs ≥3), sensitivity and specificity of c-ACE were 66.7% and 67.3%, respectively. Conclusions: In our clinical setting, the sensitivity and specificity of c-ACE for NS diagnosis were relatively poor and of little clinical utility. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Sarcoidosis is a rare chronic granulomatous systemic disorder affecting multiple organs including the nervous system (Chen and Moller, 2011). Pulmonary involvement occurs in 90% of patients, and diagnosis can be established by transbronchial or mediastinal lymph node biopsy (Israel-Biet and Valeyre, 2013). Body fluid biomarkers supporting diagnosis of systemic sarcoidosis (SS) include elevated serum angiotensinconverting enzyme (s-ACE), increased (N3.5) CD4 +/CD8 + ratio in broncho-alveolar lavage, and elevated 24 h-calciuria. Central nervous system (CNS) involvement (neurosarcoidosis, NS) occurs in 10–25% of patients (Chen and Moller, 2011) and is isolated in 1% (Vargas and Stern, 2010). NS may mimic CNS infection, neoplasia and other
Abbreviations: CNS, central nervous system; NS, neurosarcoidosis; SS, systemic sarcoidosis; ACE, angiotensin-converting enzyme; CSF, cerebrospinal fluid; c-ACE, cerebrospinalfluid angiotensin-converting enzyme; s-ACE, serum angiotensin-converting enzyme; ROC, receiver-operating characteristic; AUC, area under the curve; WBC, white blood cells. ⁎ Corresponding author at: Department of Clinical Neurosciences, Division of Neurology, University Hospital of Geneva, Rue Gabrielle-Perret-Gentil 4, 12011 Geneva, Switzerland. E-mail addresses: [email protected] (C. Bridel), [email protected] (D.S. Courvoisier), [email protected] (N. Vuilleumier), [email protected] (P.H. Lalive).
http://dx.doi.org/10.1016/j.jneuroim.2015.05.020 0165-5728/© 2015 Elsevier B.V. All rights reserved.
inflammatory disorders of the CNS such as multiple sclerosis, encephalitis and vasculitis (Reske et al., 2005). Diagnosis is challenging in cases with minimal or absent systemic involvement. Definite diagnosis requires CNS biopsy, which is not always achievable. Probable NS is defined as suggestive CNS lesions together with evidence of SS (positive systemic tissue biopsy and/or positive chest imaging together with elevated serum angiotensin converting enzyme (s-ACE)) (Zajicek et al., 1999; Marangoni et al., 2006). Cerebrospinal fluid (CSF) ACE (c-ACE) has been claimed as a valuable diagnostic tool for NS (Oksanen et al., 1985; Tahmoush et al., 2002; Reske et al., 2005). However, there is little data evaluating its performance in routine clinical practice, and results of studies are conflicting (Dale and O'Brien, 1999; Tahmoush et al., 2002). No specific biomarker is available to support NS diagnosis, underscoring the importance of c-ACE validation. In many hospitals, c-ACE is performed routinely when NS is suspected. We questioned its pertinence as a biomarker of NS, and performed a retrospective study to evaluate its clinical usefulness.
2. Materials and methods 2.1. Study design This is a retrospective monocentric chart-based study.
2
C. Bridel et al. / Journal of Neuroimmunology 285 (2015) 1–3
c-ACE levels were not significantly different in the NS group compared to the non-NS group (Table 1A). The ROC-curve (Fig. 1A) showed low diagnosis accuracy, with an AUC of 0.65. The most discriminative cut-off point was a c-ACE of 2, yielding a sensitivity of 66.7% (95% CI 0.30–0.93) and a specificity of 67.3% (95% CI 0.63–0.72) (Fig. 1B). However, no cut-off provided a good likelihood ratio. Note that given the rarity of NS in our sample (9 patients), the estimate of sensitivity was imprecise, as reflected in the large 95% CI. s-ACE levels were significantly different between groups, with a higher mean value in the NS group. However, while all 9 NS patients had values, 158 non-NS patients had missing values.
2.2. Patients Consecutive patients investigated at the Geneva University Hospital for clinically suspected NS between January 2006 and December 2012 were included. Clinical, demographical, radiological and biological information (including leucorachia, proteinorachia, intrathecal immunoglobulin synthesis, c-ACE, s-ACE), final diagnosis and, when pertinent, patho-histology, were collected. NS was considered present if (1) the patient had histologically proven NS upon CNS biopsy (definite NS), or (2) the patient had suggestive CNS lesions together with evidence of SS (probable NS). This study received approval from the local Ethics Committee.
4. Discussion
2.3. c-ACE activity measure
In this study, we find that mean c-ACE is not significantly higher in the NS group compared to the non-NS group, in contrast to what was previously reported in an American case–control study including 11 patients with probable NS and 207 controls (Tahmoush et al., 2002). Tahmoush et al. selected controls consisting of patients with other CNS and peripheral nervous system (PNS) diseases (Tahmoush et al., 2002), whereas our control group consisted of any patient with suspected NS, which is closer to a clinically relevant situation. We find that c-ACE is of poor clinical interest for NS diagnosis. Indeed, even the best cut-off value (0–2 vs. ≥ 3) only provided a sensitivity of 66.7%% with a specificity of 67.3%. This finding is actually slightly better than previously published results. Tahmoush et al. reported a sensitivity of 55% (Tahmoush et al., 2002) and Dale and O'Brien a sensitivity of 24% (Dale and O'Brien, 1999). Our study has two main limitations. First, its retrospective nature may have led to some information bias, with NS patients being investigated more closely, as exemplified by the missing data for s-ACE in non-NS patients. However, given the rarity of this condition in Western countries, a prospective study is difficult to achieve. Second, the degree of NS suspicion varies greatly among the patients included in this study.
The measure of ACE activity in CSF was assessed with the colorimetric ACE kinetic test reagents from Bühlmann Laboratories (AG, Switzerland) using the RX Daytona™ chemistry analyzer (Randox, Minnesota, USA), according to the manufacturers' instructions. Results are expressed in units per liter (U/L). Inter-series variability was 3.5% (n = 15) and the limit of quantification was 0.3 U/L. 2.4. Statistical analysis Kruskal–Wallis test was used to compare continuous variables, and Fisher exact test was used to compare categorical variables. Receiveroperating characteristic (ROC) curve and area under the curve (AUC) were performed to choose the optimal cut-off value of c-ACE and to determine the discriminative ability (sensitivity and specificity) of c-ACE as a diagnostic marker for NS. Gold standard for NS diagnosis was characteristic patho-histology of tissue biopsy. Statistical analysis was performed using STATA v13.1 (StataCorp., College Station, TX). 3. Results Of the 440 patients who underwent a lumbar puncture with c-ACE measure for suspected NS between 2006 and 2012 at the Geneva University Hospital (Table 1A), 8 had probable NS with histologicallyproven SS and 1 had definite isolated NS. Patients were close to the same age, and there were significantly more males in the NS group (Table 1A). A complete description of the 9 NS cases is found in Table 1B.
5. Conclusion In this retrospective study of 440 patients with 9 NS cases, we found that c-ACE had a relatively low sensitivity and specificity. This study does not support using c-ACE as a diagnostic laboratory test.
Table 1 Sample characteristics. A
N Age, mean (SD) Gender, N (% female) c-ACE, mean (SD) s-ACE, mean (SD)
NS
Other diagnosis
P
9 43.0 (13.4) 2 (22.2%) 4.3 (4.7) 77.2 (49.5)
431 48.5 (16.7) 261 (60.6%) 2.3 (2.7) 32.8 (18.1)
0.31 0.03 0.11 0.001
B Case
1
2
3
4
5
6
7
8
9
Age at diagnosis Gender s-ACE c-ACE CSF WBC (/mm3) CSF protein (N0.45 g/L) Intrathecal IgG synthesis Systemic sarcoidosis
56 M 9 1 6 Yes No No
48 M 86 0 3 Yes No Yes
28 M 29 3 2 No No Yes
29 F 124 3 5 Yes nd Yes
30 M 58 12 2 Yes No Yes
26 F 168 6 5 Yes nd Yes
51 M 61 4 5 Yes No Yes
68 M 105 7 nd nd nd Yes
47 M 49 1 4 No nd Yes
A. Clinical and biological characteristics of the patients included in the study. c-ACE, cerebrospinal fluid angiotensin-converting enzyme; p-ACE, plasma angiotensin-converting enzyme. B. Clinical and biological characteristics of neurosarcoidosis patients. c-ACE, cerebrospinal fluid angiotensin-converting enzyme; s-ACE, serum angiotensin-converting enzyme; CSF WBC, cerebrospinal fluid white blood cells; CSF protein, cerebrospinal fluid protein; IgG, immunoglobulin G.
C. Bridel et al. / Journal of Neuroimmunology 285 (2015) 1–3
3
A
B Cut-off value of c-ACE
Sensitivity (95% CI)
Specificity (95% CI)
Likelihood ratio
0
0.78 (0.39-0.97) 7/9 patients
0.34 (0.30-0.39) 148/431
1.18
1
0.66 (0.30-0.93) 6/9 patients
0.51 (0.46-0.56) 219/431 patients
1.35
2
0.66 (0.30-0.93) 6/9 patients
0.67 (0.63-0.72) 290/431 patients
2.00
3
0.44 (0.14-0.79) 4/9 patients
0.77 (0.73-0.81) 331/431 patients
1.91
4
0.33 (0.07-0.70) 3/9 patients
0.84 (0.80-0.87) 360/431 patients
2.07
Fig. 1. Discriminative ability of c-ACE. A. Receiver-operating characteristic (ROC) curve and area under ROC curve (AUC) of c-ACE. The ROC-curve shows low diagnosis accuracy of c-ACE, with an AUC of 0.65. B. Cut-off values of c-ACE and their associated sensitivity, specificity and positive likelihood ratio. Sensitivity, specificity and likelihood-ratio of c-ACE, according to the cut-off value of c-ACE. c-ACE, cerebrospinal fluid angiotensin-converting enzyme.
Author's contribution CB designed the study, collected the data, analyzed the data, performed the statistical analysis and wrote the manuscript. DSC checked the statistical analysis and contributed to the manuscript preparation. NV supervised the c-ACE analysis and revised the manuscript. PHL designed the study and revised the manuscript.
Competing interests Dr C. Bridel, Dr D.S. Courvoisier, DR N. Vuilleumier and Dr P.H. Lalive report no competing interests.
Acknowledgment We thank Lorraine Gibert and Manon Wehrli for their help in collecting data.
References Chen, E.S., Moller, D.R., 2011. Sarcoidosis—scientific progress and clinical challenges. Nat. Rev. Rheumatol. 7 (8), 457–467. Dale, J.C., O'Brien, J.F., 1999. Determination of angiotensin-converting enzyme levels in cerebrospinal fluid is not a useful test for the diagnosis of neurosarcoidosis. Mayo Clin. Proc. 74 (5), 535. Israel-Biet, D., Valeyre, D., 2013. Diagnosis of pulmonary sarcoidosis. Curr. Opin. Pulm. Med. 19 (5), 510–515. Marangoni, S., Argentiero, V., Tavolato, B., 2006. Neurosarcoidosis. Clinical description of 7 cases with a proposal for a new diagnostic strategy. J. Neurol. 253 (4), 488–495. Oksanen, V., et al., 1985. Angiotensin converting enzyme in cerebrospinal fluid: a new assay. Neurology 35 (8), 1220–1223. Reske, D., Petereit, H.F., Heiss, W.D., 2005. Difficulties in the differentiation of chronic inflammatory diseases of the central nervous system—value of cerebrospinal fluid analysis and immunological abnormalities in the diagnosis. Acta Neurol. Scand. 112 (4), 207–213. Tahmoush, A.J., et al., 2002. CSF-ACE activity in probable CNS neurosarcoidosis. Sarcoidosis Vasc. Diffuse Lung Dis. 19 (3), 191–197. Vargas, D.L., Stern, B.J., 2010. Neurosarcoidosis: diagnosis and management. Semin. Respir. Crit. Care Med. 31 (4), 419–427. Zajicek, J.P., et al., 1999. Central nervous system sarcoidosis—diagnosis and management. Q JM 92 (2), 103–117.