Quantification of Optic Nerve and Sheath Diameter by Transorbital Sonography: A Systematic Review and Metanalysis
Dr Katsanos has been supported by a European Academy of Neurology Research Fellowship.
ABSTRACT
BACKGROUND AND PURPOSE
To date, normal values for optic nerve diameter (OND) and optic nerve sheath diameter (ONSD) for transorbital sonography (TOS) have only been reported by individual small-scale studies, exposing a great variability in the measurement of the OND and ONSD.
METHODS
We performed a systematic review and metanalysis of available to date studies on TOS evaluation of adults without elevated intracranial pressure to provide an overview of the published literature, measuring methods and further specify normal values for OND and ONSD.
RESULTS
In total, we included 39 studies with 2,927 healthy volunteers (mean age 36.1 years, 44.4% female), so that a total of 5,854 eyes were examined. All pooled analyses were based on random effect models. Mean values for OND were provided in 13 studies. Calculated mean pooled OND value was 3.08 mm (95% confidence interval [CI], 2.9-3.25), with low heterogeneity across studies (I2 = 12.7%). Thirty-four studies provided mean values for ONSD measurement. The pool of mean ONSD measurements was 4.78 mm (95% CI, 4.63-4.94), with evidence of substantial heterogeneity between estimates ONSD (I2 = 50.6%). There were no significant differences (P = .139) in the subsequent subgroup analysis for the different geographic continents. Also, no significant differences could be recorded for the effect of age (P = .824) or gender (P = .093).
CONCLUSIONS
TOS is a frequently described and widely used method. We provide reference values of OND and ONSD that are based on metanalytical analysis. Different measuring methods of ONSD result in higher heterogeneity. Subgroup analysis revealed no significant correlation between ONSD and age, gender, or geographic origin.
Introduction
Transorbital sonography (TOS) with measurement of the optic nerve diameter (OND) and optic nerve sheath diameter (ONSD) is a basically easily feasible noninvasive diagnostic tool enabling the quantification of diameters of optic nerve and its sheaths.1 Sonographic examination of the eye was reported already in 1958 by Baum and Greenwood,2 however, limited to a purely qualitative assessment. With further technical and methodological development in the last decades, enabling the quantification of diameters of optic nerve and his sheaths, TOS became a promising novel diagnostic tool in neurology and intensive care medicine. The main areas of application are diseases with increased intracranial pressure such as idiopathic intracranial hypertension, intracranial hemorrhage, and neuro-traumatological diseases.3, 4 In this context, TOS provides a helpful diagnostic tool for early prognostication of hypoxic-ischemic encephalopathy after cardiac arrest.5 Orbital sonography can also be a helpful method for the diagnosis and follow-up of optic neuritis.6 Recently, it was shown that even optic nerve atrophy can be detected reliably by means of TOS in patients with chronic demyelinating diseases of the central nerve system, as TOS findings correlated significantly to findings of optic coherence tomography.7
With regard to the used measuring methods, there is doubtless agreement in performing the measurements of OND and ONSD 3 mm behind the optic disc.1 There are no methodological differences in literature regarding the measurement of OND, because it is consistently defined as the distance between the internal borders of the hyperechogenic area surrounding the optic nerve (Fig. 1). With regard to ONSD measurement, in the vast majority of published studies, ONSD is defined as the distance between the external borders of the hyperechogenic area surrounding the optic nerve, which represents the subarachnoidal space including the arachnoid mater.1, 5, 7 In some studies, however, the ONSD includes also the outer hypoechogenic rim, which represents the dura mater, resulting to a large heterogeneity of reported reference values.8 Several studies have further investigated the relationship between age, gender, and body-mass-index (BMI) and ONSD, but with inconsistent results.5, 7-9 In addition, the influence of ethnic/racial variability and genetic differences on the normal values has not been sufficiently investigated up to date.7, 8
In this systematic review and metanalysis, we aimed to provide the reference values, based on a metanalytical model, for OND and ONSD and their variance in an adult healthy population, give an overview of the published literature, and discuss the used measurement methods.
Methods
Literature Search and Inclusion Criteria
This metanalysis is presented according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline.10 A systemic literature search of MEDLINE (access via PUMBED) and the Scopus database was carried out by two reviewers (AHK and CS). All studies were included between January 1958 and August 2019, without language or any other restrictions. The following search terms have been used in both database searched: “optic nerve,” “sonography,” and “ultrasound.” In addition, the reference list of the identified studies was examined for further studies. In order to reduce the risk of bias, only studies with B-mode sonography and healthy adult subjects without evidence of an elevated intracerebral pressure were included.
Data Extraction and Outcome Measures
The title and abstract of the identified studies were independently reviewed by AHK and CS. Information was extracted from each included study on (1) characteristics of study participants (including mean age, sex, and country in which the study was performed); (2) data on methodology (transducer used, distance of measurement behind optic disc, and body position during measurement); and (3) healthy subjects/control group measurements in millimeters (OND and ONSD including SD and 95% confidence interval). After selection of the studies according to the criteria mentioned above, the used method was evaluated. In order to correctly evaluate the validity of the measurement methods, this procedure was performed separately by two investigators (AHK and CS). Disagreements were resolved by discussion between the two review authors and the senior author (CK). Only studies that performed the OND or ONSD measurement 3 mm behind the papilla were included. Subsequently, the method section was checked for the correct determination of OND or ONSD (see Fig. 1). The optic nerve sheath (ONS) is depicted as a thin bilateral hyperechoic line surrounding the hypoechogenic optic nerve (ON). Consequently, the correct ONSD corresponds to the distance between the outer hyperechogenic boundaries of the hyperechogenic ONS. Three studies were detected, which had measured the OND but documented it as ONSD.9, 11, 12 Studies that did not allow an exact classification were excluded from the analysis.
Statistical Analyses
Pooled estimates of mean OND and mean ONSD in both the overall and subgroup analyses were calculated with the random effects model. Meta-regression analyses of ONSD measurements with age and female prevalence were also performed under the random-effects model (method of moments). The equivalent z-test was performed for each pooled estimate, and if P < .05 it was considered statistically significant. We assessed heterogeneity among studies with the Cochran Q and I2 statistics. For all subgroup analyses, we used a standard test for heterogeneity across subgroup results to investigate for potential differences between subgroups. All analyses were conducted in STATA/MP 14.0 (StataCorp, College Station, TX) with the metan and metareg packages.
Results
Study Selection and Study Characteristics
We screened 1,253 titles and abstracts from which 149 eligible studies were retained for full-text evaluation. After careful evaluation and no disagreements between the two reviewers, 95 studies were excluded and 54 studies were selected that met the inclusion criteria (Fig. 2). Eight studies could not be included in the analysis due to a lack of standard deviation or 95% confidence interval, seven were excluded due to a lack of method description or discrepancy between method description and attached B-scan (Table 2). The range of frequencies used was between 7 and 20 MHz. Broadband ultrasound heads were used in some studies.
We included a total of 39 studies with cumulatively 2,927 patients without elevated intracranial pressure or ON disease, so that a total of 5,854 eyes were examined (Table 1). Overall 44.4% of the volunteers were female. Weighted mean patient age of all included studies was 36.1 years. Seventeen studies were conducted in Europe, seven in America, 10 in Asia, and five studies in Africa. In three studies, the OND was described either in the method description or in the B-scan instead of the ONSD, so that these measured values were subsequently handled as OND measurements.9, 11, 12
Author | Country | N | Women | Men | Females % | Mean Age | MHz | Device |
---|---|---|---|---|---|---|---|---|
Ballantyne11 | UK | 67 | 56 | 11 | 83.6% | 37 | 7 | Acuson 128 XP 10 |
Romagnuolo43 | USA | 10 | 5 | 5 | 50.0% | n.d. | 7.5 | Shimadzu SDU-400 |
Girisgin35 | Turkey | 26 | n.d. | 33.4 | 7.5 | n.d. | ||
Blehar30 | USA | 27 | 7 | 20 | 25.9% | 36.6 | 8.5 | Sonosite MicroMaxx |
Bäuerle13 | Germany | 25 | 11 | 14 | 44.0% | 45.8 | 9-3 | Philips iU22 |
Bäuerle27 | Germany | 40 | 25 | 15 | 62.5% | 37.1 | 9-3 | Philips iU22 |
Fernández-Domínguez19 | Spain | 24 | 13 | 11 | 54.2% | 48.7 | 7.5 | Toshiba Xario SSA-660a |
Bäuerle29 | Germany | 15 | 10 | 5 | 66.7% | 24.5 | 9-3 | Philips iU22 |
Anas25 | Nigeria | 400 | 208 | 192 | 52.0% | 36.9 | 7.5 | Mindray DP-8800 Plus |
Carraro15 | Italy | 37 | 27 | 10 | 73.0% | 34.1 | 7.2-14 | Toshiba Aplio XG |
Strapazzon46 | Italy | 19 | 4 | 15 | 21.1% | 39.5 | 3-11 | Esaote MyLab 30Gold |
Shirodkar45 | India | 41 | 21 | 20 | 51.2% | 27.4 | 10 | n.d. |
Aduayi23 | Nigeria | 80 | n.d. | n.d. | 7.5 | Mindray DC6 | ||
Amini24 | USA | 42 | 20 | 22 | 47.6% | 24 | 10-5 | Philips Sparq ultrasound system |
Lefferts39 | USA | 20 | 6 | 14 | 30.0% | 24 | 7.5-10 | Hitachi Prosound Alpha 7 |
Chen16 | China | 519 | 312 | 207 | 60.1% | 46.1 | 12-3 | Philips M2540A EnVisor C |
Asghar26 | Pakistan | 100 | 41 | 59 | 41.0% | 31.1 | 7.5 | Mindray (not named) |
Goeres9 | Canada | 120 | 55 | 65 | 45.8% | 29.3 | 13-6 | SonoSite M-Turbo |
Bäuerle28 | Germany | 20 | 13 | 7 | 65.0% | 47.4 | 9-3 | Philips iU22 |
Iegorova36 | Ukraine | 15 | 10 | 5 | 66.7% | n.d. | 8 | Siemens Sonoline Versa Plus |
Lochner (1) 22 | Italy | 20 | 13 | 7 | 65.0% | 46.3 | 4-11 | Toshiba Aplio XG |
Lochner (2) 21 | Italy | 21 | 14 | 7 | 66.7% | 41.7 | 7.2-14 8-15 |
Toshiba Aplio XG Philips iU22 |
Krogias20 | Germany | 15 | 5 | 10 | 33.3% | 58.3 | 7.5 | Toshiba Aplio XG |
Rehmann42 | Pakistan | 13 | 10 | 3 | 76.9% | 34.7 | n.d. | n.d. |
Tarzamni47 | Iran | 30 | 15 | 15 | 50.0% | 44 | 7.5 | Aloka Model Prosound 3500 |
Fichtner (Group 1)34 | Switzerland | 19 | 9 | 10 | 47.4% | 58 | 7-15 | Philips iU22 |
Fichtner (Group 2)34 | Switzerland | 20 | 6 | 14 | 30.0% | 51 | 7.2-14 | Toshiba Xario XG |
De Masi32 | Italy | 35 | 23 | 12 | 65.7% | 38.5 | 2.5-10 | GE Vivid 7 |
Wang12 | China | 230 | 112 | 118 | 48.7% | 43.2 | 9-3 | Philips iU22 |
Kim37 | Korea | 585 | 78 | 508 | 13.3% | 21.4 | 11-3 | GE Logiq P6 |
Toscano48 | Italy | 31 | 17 | 14 | 54.8% | 56.3 | 12 | n.d. |
Dinsmore33 | Canada | 11 | 2 | 9 | 18.2% | 33.5 | 7-13 | SonoSite Edge |
Chen17 | China | 25 | 17 | 8 | 68.0% | 47.1 | 20 | Quantel Medical Avisio |
Candeliere Merlicco14 | Spain | 36 | 16 | 10 | 44.4% | 40.1 | 7.5 | Philips HD15 Siemens Acuson X300 |
De Masi31 | Italy | 35 | 19 | 16 | 54.3% | 40.2 | 4-11 | GE Vivid 7 |
Ebraheim18 | Egypt | 30 | 30 | 0 | 100.0% | 31.7 | 7-13 | Philips iU22 |
Kishk38 | Egypt | 35 | 35 | 0 | 100.0% | 31.9 | 7-13 | Philips iU22 |
Maissan40 | Netherlands | 5 | 2 | 3 | 40.0% | 22 | n.d. | n.d. |
Omatiga41 | Nigeria | 60 | 27 | 33 | 45.0% | 52.1 | 6.5-12 | Mindray DC7 |
Sadrameli44 | USA | 24 | 24 | 0 | 100.0% | n.d. | 9 | GE Logiq E9 |
- N = Number of cases; n.d. = not documented.
Author | Country | N | Reason |
---|---|---|---|
Salgarello51 | Italy | 20 | No B-Mode sonography |
Gerling52 | Germany | 10 | Different measurement distance |
Sutherland53 | UK | 13 | Method not clearly defined |
Shah54 | USA | 20 | Method not clearly defined |
Skoloudik55 | Czech Republic | 16 | No SD or 95% CI documented |
Maude56 | Bangladesh | 136 | No SD or 95% CI documented |
Lochner57 | Italy | 21 | No SD or 95% CI documented |
Karami58 | Iran | 32 | Method not clearly defined |
Sangani59 | India | 120 | No Mean, SD, or 95% CI documented |
Lee60 | South Korea | 26 | No SD or 95% CI documented |
Lochner61 | Italy | 19 | No SD or 95% CI documented |
Thotakura62 | India | 16 | No SD or 95% CI documented |
Shrestha63 | Nepal | 27 | Method not clearly defined |
Yuzbasioglu64 | Turkey | 53 | Method not clearly defined |
Naldi65 | Italy | 40 | No SD or 95% CI documented |
- N = Number of cases; SD = Standard deviation; CI = Confidence interval.
Standard values for the OND were found in a total of 13 studies.9, 11-22 We calculated a mean pooled OND value of 3.08 mm (95% CI, 2.9-3.25 mm), with low heterogeneity across studies (I2 = 12.7%; Fig. 3). A total of 34 studies provided mean values for ONSD measurement.13, 15-18, 20-48 The pool of mean ONSD measurements was 4.78 mm (95% CI, 4.63-4.94 mm; Fig. 4), with evidence of substantial heterogeneity between estimates ONSD (I2 = 50.6%).
There were no significant differences in the subsequent subgroup analysis for the different geographic continents. Mean ONSD in Europe was 5.10 mm (95% CI, 4.78-5.41; P = .005) versus America with 4.50 mm (95% CI, 94.03-4.96; P = .985), whereas the mean ONSD in Asia was 4.68 mm (95% CI, 4.44-4.93; P = .004) and Africa 4.78 mm (95% CI, 4.63-4.94; P = .004) (P = .139; Fig. 5). No significant association of mean ONSD measurements with age (r = .0023, P = .824; Fig. 6) or gender (r = .894, P = .093; Fig. 7) could be identified.
Discussion
This is the first metanalysis of normal OND and ONSD assessed by TOS, reporting a mean OND of 3.08 mm (95% CI, 2.9-3.25) and a mean ONSD value was 4.78 mm (95% CI, 4.63-4.94) across all eligible studies. In the recent years, TOS became increasingly established as a fast, noninvasive, and easy repeatable technique in several neurological diseases, particularly with alteration of intracranial pressure.5, 13 With regard to the technical implementation of the analyzed studies, a consistent acquisition of the measured values 3 mm behind the optic disc was found. For the measurement of the OND that is consistently measured as the distance between the inner borders of the hyperechogenic rim surrounding the hypoechogenic nerve, we documented a low heterogeneity of the measured values. However, with regard to the clinically even more relevant ONSD a larger heterogeneity was found. This larger heterogeneity could be an expression of the inconsistent compliance with the correct definition of ONSD in some of the published studies, as some authors incorrectly include the hypoechogenic dura mater also into the measurement of the ONSD.49
The interrater reliability has been shown to be very high in several studies,5, 18, 27 so that for trained examiners no relevant variability effect on the measured values should occur. In over two thirds of the evaluated studies, authors used the mean value from three subsequent measurements to reduce further the effect of intrarater variability.
Several studies have investigated the relationship between age, sex, and BMI and ONSD, but with inconsistent results.5, 7-9 In our metanalysis, we did not find any association between ONSD and age or sex. However, the mean age of the examined patients was relatively young at 36.1 years, so that further investigations should be made with regard to the influence of age and possible brain atrophy. Data from children were not included in this metanalysis. From the age of 5 years on, no age-related change in ONSD (5.75 mm) could be detected in a cohort of 100 children.50 We did not perform an analysis for the BMI because it was only reported in a small percentage.
Furthermore, differences due to ethnic/racial variability and genetic differences have often been discussed in the literature as an influence on OND and ONSD.7, 8 In the subgroup analysis, sorted by geographic continent, we did not find any significant association between geographic area and ONSD in our evaluation (cf. Figure 5).
Limiting factors are the heterogeneous measurement methods of ONSD in some studies and the different technical requirements. The differences between the different ultrasound devices and settings are also a potential limiting factor. Furthermore, the lack of differences in meta-regression cannot exclude differences in individuals (ecological fallacy). Thus, we are unable to provide a specific cutoff rather than a range of normal OND and ONSD values. Moreover, the exclusion of publications without mean values, SD, and/or 95% confidence intervals and the lack of information on methods and measurement techniques in some studies are limiting factors.
In conclusion, TOS is frequently described as a widely used method. We provide reference values of OND and ONSD that are based on metanalytical analysis. Different measuring methods of ONSD result in higher heterogeneity. Subgroup analysis revealed no significant correlation between ONSD and age, gender, or geographic origin. Compared to the vast heterogeneity between ONSD measurements between different study protocols, OND values appear to be very robust and clear as evaluation revealed only low heterogeneity across studies.