Skip directly to local search Skip directly to A to Z list Skip directly to navigation Skip directly to site content Skip directly to page options
CDC Home

HuGENet™ Case Study

 

Glucocerebrosidase Gene Mutations and Parkinson Disease

 

Stuart K. Shapira, M.D., Ph.D.
National Center on Birth Defects and Developmental Disabilities
Centers for Disease Control and Prevention

line

In November, 2004, Aharon-Peretz et al. reported that mutations in the glucocerebrosidase (GBA) gene increased the susceptibility for developing Parkinson Disease.

Aharon-Peretz J et al. Mutations in the glucocerebrosidase gene and Parkinson Disease in Ashkenazi Jews.
N Engl J Med. 2004;351:1972-7.

These investigators performed a case-control study of 99 patients with Parkinson Disease, 74 patients with Alzheimer's disease, and 1543 healthy control subjects for 6 mutations in the GBA gene. All patients and controls were of Ashkenazi Jewish ancestry by family history. Parkinson Disease was diagnosed according to the United Kingdom brain-bank criteria; patients underwent a physical, neurobehavioral, and neurologic examination that incorporated the Unified Parkinson Disease Rating Scale. Alzheimer's disease patients, serving as a comparison group, met the criteria for dementia of the Alzheimer's type according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, and the criteria for probable Alzheimer's disease of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association. Healthy Ashkenazi Jews serving as a control group were from the same geographic area; the healthy controls were undergoing testing to identify heterozygosity for certain recessive diseases. The Parkinson Disease group included 55 men and 44 women, and the Alzheimer's disease group included 42 men and 32 women. No further information was provided about the cases or controls.

 

Question 1: What other information about the cases and controls would be important to know? Are there any specific concerns about the healthy controls selected for this study?

[view answer]

Question 1 - Answer

A common deficiency in case-control studies of gene-disease associations is the lack of comparability of cases and controls (i.e., controls may not come from the same population from which cases were selected). This could lead to possible selection bias and inability to generalize from the results. In this study a significant deficiency in the provided information is the age of the study groups. The healthy controls were ascertained from persons undergoing DNA mutation testing for certain recessive genetic disorders that are more prevalent in Ashkenazi Jews (Tay-Sachs disease, cystic fibrosis, etc.); most persons seeking such a test would do so because they are making reproductive plans and would like to assess their risk of having a child with one of the recessive conditions. Therefore, the healthy controls are most likely single persons planning a marriage or couples planning a pregnancy; persons in this category will be much younger than the typical person in the Parkinson Disease group. Some persons in the healthy control group may develop Parkinson Disease when they are older, and, therefore, would not be appropriate for inclusion in the control group. It would also have been helpful to know the gender distribution of the healthy controls in order to confirm that the study groups have been comparable in this respect.

[Close Window]

 

line

 

The N370S and 84GG alleles are the most frequent mutations in the GBA gene among Ashkenazi Jews, with carrier rates of 1 in 17.5 for N370S and 1 in 400 for 84GG; these alleles are associated with mild and severe Gaucher disease, respectively. Four other rare GBA mutations identified in Ashkenazi Jews are L444P, IVS2+1G-->A, V394L, and R496H. All 99 patients with Parkinson Disease, all 74 patients with Alzheimer's disease, and all 1543 healthy controls were evaluated for these 6 mutations. Among the 1543 control subjects, 95 were heterozygous for a Gaucher disease-causing mutation (6.2%). None of the control subjects were homozygous or compound heterozygous for a mutation. Of the carriers, 92 were heterozygous for N370S and 3 were heterozygous for 84GG; these results were consistent with a carrier rate of 1 in 16.7 for the N370S mutation and 1 in 514 for the 84GG mutation. 

 

Question 2: Are the observed carrier frequencies for the N370S and 84GG mutations significantly different from the carrier estimates reported in the literature?

[view answer]

Question 2 - Answer

For the N370S mutation, the observed carriers and noncarriers were 92 and 1451 (carrier frequency of 1 in 16.7). For the previously reported carrier frequency of approximately 1 in 17.5, the expected number of carriers and noncarriers in the control group would be approximately 88 and 1455, respectively. The observed and expected are compared by the Yates Chi-Square, corrected for continuity:

 

Observed

Expected

Carriers
92
88
Noncarriers
1451
1455

Chi-square = 0.05, df = 1, p = 0.8231

Thus, the observed and expected carrier frequencies for the N370S mutation are not significantly different. For the 84GG mutation, the observed carriers and noncarriers were 3 and 1540 (carrier frequency of approximately 1 in 514). For the previously reported carrier frequency of 1 in 400, the expected number of carriers and noncarriers in the control group would be approximately 3.9 and 1539.1, respectively. The observed and expected are shown in a 2 x 2 table:

 

Observed

Expected

Carriers

3
3.9

Noncarriers

1540
1539.1

Note that for a 2 x 2 table, the Yates Chi-Square test is valid only if all expected cell frequencies are equal to or greater than 5. Since this requirement is not met for the data shown here, the Pearson Chi-Square, uncorrected for continuity is calculated: Chi-square = 0.14, df = 1, p = 0.7083

The observed and expected carrier frequencies for the 84GG mutation are also not significantly different.

[Close Window]

 

line

The authors defined a “carrier" for Gaucher disease as a person who was either heterozygous for a GBA mutation, homozygous for a mutation, or a compound heterozygote for two different GBA mutations. The genotypes that were found for the “carriers” in each study group are shown:

GBA genotypes in Parkinson Disease patients and controls*

Genotype Parkinson Disease Alzheimer's Disease Healthy Controls
+/N370S
23
2
92
N370S/N370S
3
0
0
+/84GG
4
1
3
+/R496H
1
0
0
Total “Carriers”
31
3
95

*Adapted from Aharon-Peretz et al.
+ = wild type or non-mutant allele

 

Question 3: Is the definition that the authors proposed for a “carrier” for Gaucher disease appropriate for this analysis?

[view answer]

Question 3 - Answer

By grouping together homozygotes, heterozygotes, and compound heterozygotes for GBA mutations, the authors perhaps have assumed that the risk of developing Parkinson Disease for persons with each genotype is similar. However, this assumption is not valid. In fact, because the N370S mutation is considered a “mild” mutation, while the 84GG is considered a “severe” mutation, carriers for these mutations may have quite different risks for developing Parkinson Disease, depending upon their genotypes. In addition, even though the proportion of homozygous persons in the Parkinson Disease group is only 9.7% (3 N370S homozygotes divided by 31 total “carriers”), the risk for homozygotes or compound heterozygotes to develop Parkinson Disease may be much higher than among simple heterozygotes, which could influence the overall risk assessment for “carriers.” Therefore, by grouping all genotypes with GBA mutations into a single category it is not possible to determine if the genotype-specific risks for developing Parkinson Disease are different, each with perhaps quite different clinical and public health consequences.

[Close Window]

 

line

The rates of “carriage of Gaucher disease” (heterozygotes plus homozygotes for any mutation) were calculated for patients with Parkinson Disease, Alzheimer's disease, and control subjects:

Rates of carriage of GBA mutations among patients with Parkinson Disease and Controls*

Patient Populations Number Tested Number of “Carriers” (%) 95% Confidence Interval
Parkinson Disease
99
31 (31.3%)
22.2-40.4
Alzheimer's Disease
74
3 (4.1%)
0.0-8.5
Healthy Controls
1543
95 (6.2%)
5.0-7.4

*Adapted from Aharon-Peretz et al.

 

Question 4: Based on these findings, what is the odds ratio (estimated relative risk) for Parkinson Disease among “carriers” of Gaucher disease? Is the rate of “carriage” of a GBA mutation among patients with Alzheimer's disease increased compared with healthy controls?

[view answer]

Question 4 - Answer

Odds ratios are calculated from the 2 x 2 tables as follows:

  Parkinson Disease Alzheimer's Disease Odds Ratio Parkinson Disease Healthy Controls Odds Ratio
+/+
68
71
ref
68
1448
ref
“Carrier”
31
3
10.8
31
95
6.9
Total
99
74
 
99
1543
 

The authors concluded that patients with Parkinson Disease had significantly greater odds for being “carriers” of Gaucher disease than did patients with Alzheimer's disease (odds ratio, 10.8; 95% confidence interval, 3.0-46.6; p<0.001) or control subjects (odds ratio, 6.9; 95% confidence interval, 4.2-11.4; p<0.001).

The odds ratio for GBA mutation in Alzheimer's disease was:

  Alzheimer's Disease Healthy Controls Odds Ratio
+/+
71
1448
ref
“Carrier”
3
95
0.6
Total
74
1543
 

At first glance, finding an odds ratio less than one might suggest that “carriers” for a GBA gene mutation have a lower risk for developing Alzheimer's disease. However, the rate of “carriage” of a GBA gene mutation among patients with Alzheimer's disease did not differ significantly from that in healthy controls (95% confidence interval 0.2-2.2, p=0.62).

[Close Window]

 

line

The authors' overall conclusions from the study were (1) the prevalence of GBA mutations in the population of Ashkenazi Jews with Parkinson Disease by far outweighs the reported prevalence of mutations in other susceptibility genes, such as parkin and SNCA, and (2) mutations in the GBA gene emerge as a strong genetic determinant predisposing people to Parkinson Disease.

 

Question 5: What additional analysis of the data should be performed in order to determine the validity of the authors' second conclusion?

[view answer]

Question 5 - Answer

The authors' second conclusion is not rigorously supported unless an analysis is performed to evaluate the effects of specific alleles, or the contribution of homozygosity for a mutant allele versus heterozygosity, on the risk for developing Parkinson Disease in this patient population. Although the authors chose not to do this analysis, perhaps because of small sample sizes, a determination of the odds ratios for developing Parkinson Disease among persons with each GBA genotype should be performed since the results will be helpful for determining the risk for persons with each genotype of developing Parkinson Disease, relative to the prevalence of Parkinson Disease in the population.

[Close Window]

 

line

In order to determine the GBA genotypic contribution toward the development of Parkinson Disease, the Parkinson Disease and healthy control groups are compared as shown:

Genotype Parkinson Disease Healthy Controls OR 95% Confidence Interval p value
+/+
68
1448
ref
   
+/N370S
23
92
5.32
3.19-8.90
< 0.001
N370S/N370S
3
0
     
+/84GG
4
3
28.39
6.96-115.65
< 0.001
Total
98
1543
     

 

Question 6: Since there were no N370S homozygotes among the healthy controls, how can the odds ratio be calculated for this genotype?

[view answer]

Question 6 - Answer

One way to obtain the expected number of N370S homozygotes among the controls is to assume that the GBA alleles in the control population are in Hardy-Weinberg equilibrium. The calculated expected number of controls that are homozygous for the N370S mutation is 1.3713, which is then used to estimate the odds ratio.

[Close Window]

 

line

The observed number of healthy controls with each genotype is compared to the expected number of controls, based on Hardy-Weinberg equilibrium:

Genotype Observed Expected
+/+
1448
1449.4622
+/N370S
92
89.1678
N370S/N370S
0
1.3713
+/84GG
3
2.9076
N370S/84GG
0
0.0894
84GG/84GG
0
0.0014

 

Question 7: Are the observed numbers for the genotypes that comprise the healthy controls significantly different from Hardy-Weinberg equilibrium? Is it valid to have used 1.3713 as the number of healthy controls with the N370S/N370S genotype in order to calculate the odds ratio?

[view answer]

Question 7 - Answer

The numbers of observed and expected controls with each genotype can be compared by the Chi-square test. The result for these data is: Chi-square = 1.5566; df = 3; p = 0.500-0.750. Based on Hardy-Weinberg equilibrium, the observed frequencies for each genotype in the healthy controls do not differ significantly from those expected for all three alleles: +, N370S, and 84GG. Therefore, it is valid to utilize the expected number of controls for the N370S/N370S genotype in order to calculate the odds ratio.

[Close Window]

 

line

The odds ratios for developing Parkinson Disease, the 95% confidence intervals, and the p values are shown for each GBA gene genotype:

Genotype Parkinson Disease Healthy Controls OR 95% Confidence Interval p value
+/+
68
1448
ref
   
+/N370S
23
92
5.32
3.19-8.90
< 0.001
N370S/N370S
3
0
46.59
7.74-279.09
< 0.001
+/84GG
4
3
28.39
6.96-115.65
< 0.001
Total
98
1543
     

 

Question 8: How strong is the association between each of the GBA gene genotypes and Parkinson Disease?

[view answer]

Question 8 - Answer

All calculated odds ratios were statistically significant, with p values < 0.001. The analyses showed that (1) N370S homozygotes were approximately 47 times more likely than people without this genotype to have Parkinson Disease, but accounted for only 3 of 98 cases, (2) N370S heterozygotes were approximately 5 times more likely to have Parkinson Disease, and (3) 84GG heterozygotes were approximately 28 times more likely to have Parkinson Disease, but accounted for only 4 out of 98 cases. Therefore, of all of the “carrier” classes defined by the authors, the most important contribution at the population level to risk for developing Parkinson Disease comes from the N370S heterozygotes which are 5 times more likely to have Parkinson Disease.

[Close Window]

 

line

In order to quantify the contribution of the various GBA gene genotypes to the occurrence of Parkinson Disease at the population level, one can calculate the associated attributable fraction utilizing the formula of Miettinen:

Attributable fraction = fc (R – 1) / R

Where fc is the fraction of cases with the risk factor and R is the measure of relative risk (or odds ratio for rare diseases). The following table summarizes the calculation of attributable fraction for each genotype:

Genotype fc R Attributable Fraction
+/N370S
23/98
5.32
19.1%
N370S/N370S
3/98
46.59
3.0%
+/84GG
4/98
28.39
3.9%

 

In order to estimate the attributable fraction from these data, it was necessary to assume that the cases and controls were comparable and population-based.


Question 9: Are the results of this analysis valid? If the result is correct, what is then suggested about the role of the GBA gene genotypes in Parkinson Disease at the population level?

[view answer]

Question 9 - Answer

The Parkinson Disease and healthy control study groups were all of Ashkenazi Jewish heritage and presumably sampled from the same population (by referral to a tertiary medical center for northern Israel ). To estimate attributable fraction for the Ashkenazi Jewish population, we would have to assume that the cases are representative of all Ashkenazi Jewish persons with Parkinson Disease and that the controls are also a representative (unbiased) sample of the Ashkenazi Jewish population. The authors did not present the age distribution of the Parkinson Disease patients or the age or gender distribution of healthy controls, who were probably much younger on average than the Parkinson Disease patients, a factor that would tend to bias the study against finding an association with GBA genotype. If results of the analysis are correct, homozygosity for N370S and heterozygosity for 84GG are strong risk factors for Parkinson Disease but account for only a small proportion of cases. Heterozygotes for N370S were much more prevalent than homozygotes among cases, with an attributable fraction of almost 20%. Although all the studied GBA genotypes taken together accounted for 26% of the Parkinson Disease cases, most people with Parkinson Disease in the study population did not have one of the 6 analyzed mutations in the GBA gene.

[Close Window]

 

line

The estimated absolute risk (incidence) of Parkinson Disease by genotype can be determined using the estimated frequencies of each genotype, and summing the genotype specific incidence rates in order to calculate the absolute risk. The allele and genotype frequencies are calculated as follows (allele frequencies derived from the population genetics expressions of p + q + r = 1 and [(p x p) + (2p x q) + (q x q) +

(2p x r) + (2q x r) + (r x r) = 1]; these formulas assume that both the N370S and 84GG alleles are in Hardy-Weinberg equilibrium, which was shown previously to be the case):

Allele Frequency Mutation N370S 84GG None (+)
.02981
N370S
.00089
.00003
.02889
.00097
84GG
.00003
.0000009
.00094
.96922
None (+)
.02889
.00094
.93939

+ = wild type or non-mutant allele

Genotype Frequencies:
+/+ = 0.93939
+/N370S = 0.02889 x 2 = 0.05778
+/84GG = 0.00094 x 2 = 0.00188
N370S/N370S = 0.00089

 

Question 10: Based on these genotype frequencies, and assuming that the total population risk of Parkinson Disease is 1 per 100 people (1%), what is the estimated absolute risk of Parkinson Disease by genotype?

[view answer]

Question 10 - Answer

Assume that the odds ratio is an approximation of the relative risk, and that the total population risk is 0.01. The population wide incidence (sum of the genotype specific incidence rates) and the absolute risk are calculated by summing the products of the genotype frequencies and the odds ratios, where “x” is the absolute risk of Parkinson Disease in people without any of the GBA gene mutations (wild type) :

+/+
(0.9393)(1)(x)
0.9393x
+/N370S
(0.05778)(5.32)(x)
0.3074x
+/84GG
(0.00188)(28.39)(x)
0.0544x
N370S/N370S
(0.00089)(46.59)(x)
0.0415x
Sum:
1.3426x

Prevalence = 0.01 = 1.3426x
x = 0.00744

The risk for developing Parkinson Disease for persons with each genotype is then determined by multiplying the baseline absolute risk (x) by the odds ratio:

Genotype Incidence/1000 Risk of Developing Parkinson Disease With the Genotype
+/+
(1)(.00744)(1000)
0.7%
+/N370S
(5.32)(.00744)(1000)
4.0%
+/84GG
(28.39)(.00744)(1000)
21.1%
N370S/N370S
(46.59)(.00744)(1000)
34.7%
[Close Window]

 

line

The calculation of absolute risk of Parkinson Disease by genotype showed that there are significant risks for developing Parkinson Disease, over the baseline population risk of 1%, for both N370S and 84GG carriers (4.0% and 21.1%, respectively), as well as for N370S homozygotes (34.7%). However, it is crucial to determine whether these are important risks from a population-based public health perspective.

 

Question 11: Would it beneficial to perform population-based screening of Ashkenazi Jews for the N370S mutation--the mutation with the highest incidence in the population--in order to identify those persons with an increased risk of developing Parkinson Disease?

[view answer]

Question 11 - Answer

Screening 1000 Ashkenazi Jews will identify approximately 59 persons who are N370S heterozygotes or homozygotes:

(1000)(0.05778) + (1000)(.00089)
57.8 + 0.9 = approximately 59 persons

Based on the risks that persons with each genotype will develop Parkinson Disease, of the 59 identified persons, approximately 3 would be expected to develop the condition:

(1000)(.05778)(0.040) + (1000)(.00089)(0.347)
2.31 + 0.31 = approximately 3 persons

Therefore, 95% of identified persons with the N370S mutation are not expected to develop Parkinson Disease. This finding is in contrast to the authors' conclusion that “mutations in the GBA gene emerge as a strong genetic determinant predisposing people to Parkinson Disease.” This statement may relate to the fact that mutations in the GBA gene, particularly in the Ashkenazi Jewish population, are much more common than mutations in other genes implicated in the genetic susceptibility to Parkinson Disease (parkin, SNCA, SNCAIP, etc.). However, the individual risk of developing Parkinson Disease for carriers of a mutation in the GBA gene (particularly the N370S mutation), is still very low. Because there is no way to distinguish the N370S heterozygotes or homozygotes who will develop Parkinson Disease from those who won't, and there are no clear lifestyle modifications or medical interventions to lower the risk, N370S carrier identification would cause needless anxiety about the uncertain likelihood of developing a major future health problem without providing a clear health benefit.

[Close Window]

line

View all answers.

line

 

Contact Us:
  • CDC-INFO Contact Center
    Phone: 800-232-4636 (800-CDC-INFO)

    Hearing Impaired: 888-232-6348
  • cdcinfo@cdc.gov
  • Additional information for Public Health Genomics is available on our contact page.
USA.gov: The U.S. Government's Official Web PortalDepartment of Health and Human Services
Centers for Disease Control and Prevention   1600 Clifton Rd. Atlanta, GA 30333, USA
800-CDC-INFO (800-232-4636) TTY: (888) 232-6348 - Contact CDC–INFO
A-Z Index
  1. A
  2. B
  3. C
  4. D
  5. E
  6. F
  7. G
  8. H
  9. I
  10. J
  11. K
  12. L
  13. M
  14. N
  15. O
  16. P
  17. Q
  18. R
  19. S
  20. T
  21. U
  22. V
  23. W
  24. X
  25. Y
  26. Z
  27. #