Plasma Globotriaosylsphingosine and α-Galactosidase A Activity as a Combined Screening Biomarker for Fabry Disease in a Large Japanese Cohort

Round 1

Reviewer 1 Report

This study is promising if the lyso-Gb3 can distinguish FD in different types.

To demonstrate the effectiveness of the screening methods, sensitivity, specificity, predictive values are often used, but no information was provided. Only ROC analysis was conducted but it is not enough to show the effectiveness.

In the GLA genetic analyses, significant proportions of patients declined. It can affect the results. Authors should indicate lyso-Gb3 and GLA activity levels in Fig 4.

Authors stated statistical methods, but actual method applied was unclear.

For each dataset, statistical method used should be specified. What data followed normal distribution? In small number dataset, any statistical test cannot be applied.

Author Response

Comments from Reviewer #1 and our response

Major comment 1. To demonstrate the effectiveness of the screening methods, sensitivity, specificity, predictive values are often used, but no information was provided. Only ROC analysis was conducted but it is not enough to show the effectiveness.

You are correct. Unfortunately, not all of the participants underwent genetic analysis, so it was difficult to test for sensitivity and specificity.

Major comment 2. In the GLA genetic analyses, significant proportions of patients declined. It can affect the results. Authors should indicate lyso-Gb3 and GLA activity levels in Fig 4.

According to your comment, we added the lyso-Gb3 and GLA activity levels of patients whose GLA analysis was declined in Figure 4.

And for the Figure 4 legend (A and B), we rephrased “(A and B) Male patients enrolled between July 1, 2014 and December 31, 2016 belonging to control, classic, late-onset, late-onset biopsy-proven, or class 2 groups; patients belonging to the declined group (n = 17) and not detected group (n = 1) summarized in flowcharts in Figure 2 were excluded from the analysis.” as “(A and B) Male patients enrolled between July 1, 2014 and December 31, 2016 belonging to control, classic, late-onset, late-onset biopsy-proven, class 2, or declined groups; patients belonging to late-onset biopsy-proven group, which was unsuitable for statistical analysis due to its small number, and not detected group (n = 1) summarized in flowcharts in Figure 2, and declined group, in which GLA variants were not analyzed, were excluded from the analysis.”

For the Figure 4 legend (C and D), we rephrased “(C and D) Female patients enrolled between July 1, 2014 and December 31, 2016 belonging to control, classic, late-onset, late-onset biopsy-proven, or class 2 groups; patients belonging to late-onset biopsy-proven (n = 1) and declined group (n = 22) summarized in flowcharts in Figure 3 were excluded from the analysis.” as “(C and D) Female patients enrolled between July 1, 2014 and December 31, 2016 belonging to control, classic, late-onset, late-onset biopsy-proven, class 2, or declined groups; patients belonging to late-onset group, which was unsuitable for statistical analysis due to its small number, and declined group, in which GLA variants were not analyzed, were excluded from the analysis.”

Major comment 3. Authors stated statistical methods, but actual method applied was unclear.

For each dataset, statistical method used should be specified. Statistical analysis and graph preparation. What data followed normal distribution?

2.9. Statistical analyses

According to your comment on the method of statistical analysis, we rephrased “The data distributions were examined using the Shapiro–Wilk test for n ≤ 2,000 or using the Kolmogorov–Smirnov–Lillefors test for n > 2,000. Normally distributed data were assessed for variance using the F test. Statistical analyses—Student’s t test, Welch’s t test, and Wilcoxon rank-sum test as appropriate—were performed using JMP12 (SAS In-stitute, Cary, NC, USA) and graphed in SigmaPlot 14 (Systat Software, San Jose, CA, USA). Values were considered statistically significant at P < 0.05. The ROC curve and the AUC were analyzed and generated in SigmaPlot 14.” as “The data distributions were examined using the Shapiro–Wilk test for n ≤ 2,000 or using the Kolmogorov–Smirnov–Lillefors test for n > 2,000 to determine whether they showed a normal distribution. Normally distributed data were assessed for variance using the F test. A Student’s t-test was used for comparisons between two unpaired groups with homogeneous variances. Welch’s t-tests were applied for comparisons between two unpaired groups with heterogeneous variances. The Wilcoxon rank-sum test was used for comparison between two unpaired groups when the data were not normally distributed. Values were considered statistically significant at P < 0.05. Data were statistically analyzed in JMP®12 (SAS Institute, Cary, NC, USA) and graphed in SigmaPlot 14 (Systat Software, San Jose, CA, USA). Receiver operating characteristic (ROC) curve and the area under the curve (AUC) was analyzed and generated in SigmaPlot 14.”

In small number dataset, any statistical test cannot be applied.

According to your comment, we deleted the results of statistical test of Figure 4A Biopsy-proven (n = 1), 4B Biopsy-proven (n = 1), 4C Late-onset (n = 2), and 4D Late-onset (n = 2).

Reviewer 2 Report

This is a well written and  comprehensive manuscript

This reviewer suggests acceptance pending:

• minor typographical proof reading
• further details and information on subjects and their selection
• more detailed description of study protocol

Author Response

Comments from Reviewer #2 and our response

Major comment 1. minor typographical proof reading

Title

We rephrased “Plasma Globotriaosylsphingosine and α-Galactosidase Activity as a Combined Screening Biomarker for Fabry Disease in a Large Japanese Cohort” as “Plasma Globotriaosylsphingosine and α-Galactosidase A Activity as a Combined Screening Biomarker for Fabry Disease in a Large Japanese Cohort.”

Materials and Methods

2.8. Defining FD

(Smid et al., 2014) should read [11].

2.9. Statistical analyses

We rephrased “The ROC curve and the AUC were analyzed and generated in SigmaPlot 14.” as “Receiver operating characteristic (ROC) curve and the area under the curve (AUC) was analyzed and generated in SigmaPlot 14.”

Results

3.9. Receiver operating characteristic (ROC) curve analysis for plasma lyso-Gb3 levels, GLA activity, and GLA/lyso-Gb3 ratio

We rephrased “3.9. Receiver operating characteristic (ROC) curve analysis for plasma lyso-Gb3 levels, GLA activity, and GLA/lyso-Gb3 ratio” as “3.9. ROC curve analysis for plasma lyso-Gb3 levels, GLA activity, and GLA/lyso-Gb3 ratio”

We rephrased “To compare the ROC curves and area under the curve (AUC) values of these biomarkers in the same graph, instead of lyso-Gb3 levels, we used the reciprocal (1/lyso-Gb3 levels), which are low in FD probands.” as “To compare the ROC curves and AUC values of these biomarkers in the same graph, instead of lyso-Gb3 levels, we used the reciprocal (1/lyso-Gb3 levels), which are low in FD probands.”

References

18. den Dunnen, J.T.; Dalgleish, R.; Maglott, D.R.; Hart, R.K.; Greenblatt, M.S.; McGowan-Jordan, J.; Roux, A.F.; Smith, T.; Anto-narakis, S.E.; Taschner, P.E. HGVS recommendations for the description of sequence variants: 2016 update. Hum Mutat 2016, 37, 564–569, doi: 10.1002/humu.22981

We put the period at the last of this reference in revised manuscript.

27. Shimotori, M.; Maruyama, H.; Nakamura, G.; Suyama, T.; Sakamoto, F.; Itoh, M.; Miyabayashi, S.; Ohnishi, T.; Sakai, N.; Wataya-Kaneda, M.; Kubota, M.; Takahashi, T.; Mori, T.; Tamura, K.; Kageyama, S.; Shio, N.; Maeba, T.; Yahagi, H.; Tanaka, M.; Oka, M.; Sugiyama, H.; Sugawara, T.; Mori, N.; Tsukamoto, H.; Tamagaki, K.; Tanda, S.; Suzuki, Y.; Shinonaga, C.; Miya-zaki, J.; Ishii, S.; Gejyo, F. Novel mutations of the GLA gene in Japanese patients with Fabry disease and their functional characterization by active site specific chaperone. Hum Mutat 2008, 29, 331, doi: 10.1002/humu.9520.

We corrected the name of Journal by italic type in revised manuscript.

Major comment 2. further details and information on subjects and their selection more detailed description of study protocol

Materials and Methods

2.2. Patient enrollment

According to your comment, we rephrased “Clinical evaluations included: cardiac evaluation, including an electrocardiogram, echocardiogram, and cardiac magnetic resonance imaging (for unexplained left ventricular hypertrophy or unexplained cardiac failure); nephrological evaluation with biochemical examination, urinalysis, imaging, and kidney biopsy (for chronic kidney disease, unexplained proteinuria, or pathological findings consistent with FD); neurological evaluation with magnetic resonance imaging (early-onset stroke or transient ischemic attack); and pediatric evaluation in children with early-onset classic manifestations [7]. Patients with known FD and their relatives were excluded.” as “Clinical evaluations included: cardiac evaluation, including an electrocardiogram, echocardiogram, and cardiac magnetic resonance imaging (for unexplained left ventricular hypertrophy or unexplained cardiac failure); nephrological evaluation with biochemical examination, urinalysis, imaging, and kidney biopsy (for chronic kidney disease, unexplained proteinuria, or pathological findings consistent with FD); neurological evaluation with magnetic resonance imaging (early-onset stroke or transient ischemic attack); and pediatric evaluation in children with early-onset classic manifestations—acroparesthesia, clustered angiokeratoma, cornea verticillata, and hypohidrosis [7]. Acroparesthesia is defined as pain in the hands and/or feet, with an onset of pain in childhood or adolescence, and/or a course characterized by exacerbations that are provoked by fever, exercise, or heat, as well as a decreased cold sensation [11]. Clustered angiokeratoma should be present in the bathing trunk, periumbilical, and/or perioral regions. Cornea verticillata should be evaluated using a slit lamp, in the absence of amphiphilic drug use. Hypohidrosis is defined as low or no sweating even in an environment (high temperature or humidity) that encourages sweating [11]. Patients with known FD and their relatives were excluded.”

Reviewer 3 Report

Summary: The manuscript titled “Plasma Globotriaosylsphingosine and α-Galactosidase Activity as a Combined Screening Biomarker for Fabry Disease in a Large Japanese Cohort” is a research article in which the authors report the final results of a complete analysis of a patient with potential Fabry disease evaluating plasma globotriaosylsphingosine and GLA activity as a combined screening marker. The study includes a screening of 5,691 patients and measuring the abovementioned parameters in their plasma, which provides invaluable evidence for the potential of future utilization of plasma globotriaosylsphingosine and GLA activity combination measurements as a primary screening biomarker in clinical diagnosis for classic, late-onset, and late-onset biopsy-proven Fabry disease probands. This novel approach will further strengthen and facilitate the current diagnostic evaluation of patients with Fabry disease and, therefore, will be of great interest to scientific readers.

The manuscript is well written, easy to follow, and understand with clearly labeled figures with robust data provided and concise, straightforward content. The study is well designed, methods and statistical analyses are appropriate and up to date. The reference list is comprehensive and complete. Results are clearly explained and support the conclusions of the authors. I do not find any significant concerns or deficiencies with the current manuscript; however, a typos check is warranted for identifying minor errors such as double periods or missing ones, etc. In addition, I would strongly encourage the authors to add the following to the methods section:

• To “2. Patient enrollment” - add a statement for IRB protocol approval for patient enrollment in the study.
• To “2.3. Sample Collection” - add a description of plasma preparation procedure, storage, and handling accordingly to the assay in which it was used.

Author Response

Comments from Reviewer #3 and our response

Major comment 1. a typos check is warranted for identifying minor errors such as double periods or missing ones, etc.

Title

We rephrased “Plasma Globotriaosylsphingosine and α-Galactosidase Activity as a Combined Screening Biomarker for Fabry Disease in a Large Japanese Cohort” as “Plasma Globotriaosylsphingosine and α-Galactosidase A Activity as a Combined Screening Biomarker for Fabry Disease in a Large Japanese Cohort.”

Materials and Methods

2.8. Defining FD

(Smid et al., 2014) should read [11].

2.9. Statistical analyses

We rephrased “The ROC curve and the AUC were analyzed and generated in SigmaPlot 14.” as “Receiver operating characteristic (ROC) curve and the area under the curve (AUC) was analyzed and generated in SigmaPlot 14.”

Results

3.9. Receiver operating characteristic (ROC) curve analysis for plasma lyso-Gb3 levels, GLA activity, and GLA/lyso-Gb3 ratio

We rephrased “3.9. Receiver operating characteristic (ROC) curve analysis for plasma lyso-Gb3 levels, GLA activity, and GLA/lyso-Gb3 ratio” as “3.9. ROC curve analysis for plasma lyso-Gb3 levels, GLA activity, and GLA/lyso-Gb3 ratio”

We rephrased “To compare the ROC curves and area under the curve (AUC) values of these biomarkers in the same graph, instead of lyso-Gb3 levels, we used the reciprocal (1/lyso-Gb3 levels), which are low in FD probands.” as “To compare the ROC curves and AUC values of these biomarkers in the same graph, instead of lyso-Gb3 levels, we used the reciprocal (1/lyso-Gb3 levels), which are low in FD probands.”

References

18. den Dunnen, J.T.; Dalgleish, R.; Maglott, D.R.; Hart, R.K.; Greenblatt, M.S.; McGowan-Jordan, J.; Roux, A.F.; Smith, T.; Anto-narakis, S.E.; Taschner, P.E. HGVS recommendations for the description of sequence variants: 2016 update. Hum Mutat 2016, 37, 564–569, doi: 10.1002/humu.22981

We put the period at the last of this reference in revised manuscript.

27. Shimotori, M.; Maruyama, H.; Nakamura, G.; Suyama, T.; Sakamoto, F.; Itoh, M.; Miyabayashi, S.; Ohnishi, T.; Sakai, N.; Wataya-Kaneda, M.; Kubota, M.; Takahashi, T.; Mori, T.; Tamura, K.; Kageyama, S.; Shio, N.; Maeba, T.; Yahagi, H.; Tanaka, M.; Oka, M.; Sugiyama, H.; Sugawara, T.; Mori, N.; Tsukamoto, H.; Tamagaki, K.; Tanda, S.; Suzuki, Y.; Shinonaga, C.; Miya-zaki, J.; Ishii, S.; Gejyo, F. Novel mutations of the GLA gene in Japanese patients with Fabry disease and their functional characterization by active site specific chaperone. Hum Mutat 2008, 29, 331, doi: 10.1002/humu.9520.

We corrected the name of Journal by italic type in revised manuscript.

Major comment 2. In addition, I would strongly encourage the authors to add the following to the methods section:

To “2. Patient enrollment” - add a statement for IRB protocol approval for patient enrollment in the study.

According to your comment, we added the following sentence in patient enrollment “The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Niigata University School of Medicine (protocol code 1802, 2367, H25–661-627, H27–805-767).”

To “2.3. Sample Collection” - add a description of plasma preparation procedure, storage, and handling accordingly to the assay in which it was used.

According to your comment, we added a description of plasma preparation procedure, storage, and handling accordingly to the assay in which it was used.

“Blood specimens for the primary screening were collected in Venoject II collection tubes (Terumo, Tokyo, Japan) and immediately refrigerated. Plasma specimens were obtained by centrifugation of whole blood at 1,670 × g for 10 min using a KS-5200C centrifuge (Kubota, Tokyo, Japan) in a refrigerated room at 4°C and stored frozen at −20°C until measurement.”

Round 2

Reviewer 1 Report

In sections 3.2 and 3.3, statistical methods were not specified. Only p values do not provide rationale of the analysis.

Author Response

Comments from Reviewer #1 and our response

Major comment 1. In sections 3.2 and 3.3, statistical methods were not specified. Only p values do not provide rationale of the analysis.

3.2. Plasma lyso-Gb3 levels

We rephrased “Based on plasma lyso-Gb3 levels, patients were classified as positive (≥2.0 ng/mL) or negative (<2.0 ng/mL). The median lyso-Gb3 levels were 14.6 ng/mL [interquartile range (IQR) 4.1–173.5] in the 14 males in the positive group and 0.4 ng/mL (IQR 0.2–0.6) in the 3,425 males in the negative group; this difference was significant (P < 0.0001). The median lyso-Gb3 levels were 15.0 ng/mL (IQR 7.8–21.8) for the 19 females in the positive group and 0.4 ng/mL (IQR 0.3–0.6) for the 2,233 females in the negative group; this difference was also significant (P < 0.0001).” as “Based on plasma lyso-Gb3 levels, patients were classified as positive (≥2.0 ng/mL) or negative (<2.0 ng/mL). The median lyso-Gb3 levels were 14.6 ng/mL [interquartile range (IQR) 4.1–173.5] in the 14 males in the positive group and 0.4 ng/mL (IQR 0.2–0.6) in the 3,425 males in the negative group; this difference was significant (Wilcoxon rank-sum test, P < 0.0001). The median lyso-Gb3 levels were 15.0 ng/mL (IQR 7.8–21.8) for the 19 females in the positive group and 0.4 ng/mL (IQR 0.3–0.6) for the 2,233 females in the negative group; this difference was also significant (Wilcoxon rank-sum test, P < 0.0001).”

3.3. Plasma GLA activity

We rephrased “The median plasma GLA activity in the 14 males in the positive group was 0.9 nmol/h/mL (IQR 0.4–1.6), which was significantly lower (P < 0.0001) than the median of 8.1 nmol/h/mL (IQR 6.8–9.6) in the 3,425 males in the negative group. GLA activity was low in 13 males in the positive group and in 26 males in the negative group. One male in the positive group had normal GLA activity. The median GLA activity in the 19 females in the positive group was 3.0 nmol/h/mL (IQR 1.9–5.9), which was significantly lower (P < 0.0001) than that in the 2,233 females in the negative group of 7.4 nmol/h/mL (IQR 6.2–8.8). Eight females in the positive group had normal GLA activity and 21 females in the negative group had low GLA activity.” as “The median plasma GLA activity in the 14 males in the positive group was 0.9 nmol/h/mL (IQR 0.4–1.6), which was significantly lower (Wilcoxon rank-sum test, P < 0.0001) than the median of 8.1 nmol/h/mL (IQR 6.8–9.6) in the 3,425 males in the negative group. GLA activity was low in 13 males in the positive group and in 26 males in the negative group. One male in the positive group had normal GLA activity. The median GLA activity in the 19 females in the positive group was 3.0 nmol/h/mL (IQR 1.9–5.9), which was significantly lower (Wilcoxon rank-sum test, P < 0.0001) than that in the 2,233 females in the negative group of 7.4 nmol/h/mL (IQR 6.2–8.8). Eight females in the positive group had normal GLA activity and 21 females in the negative group had low GLA activity.”