3.1 Analysis framework and flow chart
3.2 Instrumental variable screening
According to the screening criteria for instrumental variables in this study, SNPs with linkage disequilibrium were excluded. SNPs associated with to anemia indicators were included as instrumental variables after matching with the GWAS data of AR. The number of instrumental variables for anemia is presented in Table 1. The MR analysis yielded significant results (p < 0.05), and the F-test statistics for the instrumental variables of anemia were all greater than 10. This indicated that the majority of the SNPs screened in this study were strong instrumental variables, and any potential bias due to weak instrumental variables was limited.
Table 1
Screening of instrumental variables for anemia and allergic rhinitis and F test results on the strength of instrumental variables
Exposure
|
Number of SNPs
|
Median of F
|
Minimum of F
|
Maximum of F
|
Other and unspecified anemias
|
10
|
22.92363
|
32.79645
|
20.91405
|
SNPs, single nucleotide polymorphisms; F, F-statistics.
3.3 MR causal effect estimates
We employed MR-Egger, weighted median, IVW, simple mode (SM) and weighted mode models for the analysis. Scatter plots of the SNP effect estimates are shown in Fig. 2. It is apparent that the fitting curves of the scatter plots for all five models exhibit consistent directions, with relatively consistent slopes observed across most models. Additionally, the intercept of the IVW model closely approximates 0.
The estimation results of the causal effect of anemia on the incidence of AR are shown in Table 2. The results of the IVW model indicate a significant causal association between anemia and AR. Specifically, more severe anemia was associated with a higher risk of AR (OR = 1.02, p < 0.05).
Table 2
Mendelian randomization causal effect estimates of anemia on the onset of allergic rhinitis
Exposure
|
Method
|
nsnp
|
Number of SNPs
|
Beta
|
Standard error
|
OR95CI
|
p-value
|
Other and unspecified anemias
|
Inverse variance weighted
|
10
|
0.020549
|
0.010396
|
1.020762
|
1.020762 (1.000173, 1.041775)
|
0.048088
|
Other and unspecified anemias
|
MR Egger
|
10
|
0.000776
|
0.026513
|
1.000776
|
0.977368 (0.950099, 1.054156)
|
0.977368
|
Other and unspecified anemias
|
Simple mode
|
10
|
0.01153
|
0.012941
|
1.011596
|
0.396149 (0.986261, 1.037583)
|
0.396149
|
Other and unspecified anemias
|
Weighted median
|
10
|
0.012099
|
0.010181
|
1.012172
|
1.012172 (0.992176, 1.032572)
|
0.234659
|
Other and unspecified anemias
|
Weighted mode
|
10
|
0.01153
|
0.012018
|
1.011596
|
0.36243 (0.988047, 1.032572)
|
0.36243
|
SNP, single nucleotide polymorphism; OR, odds ratio; CI, confidence interval.
3.4 Sensitivity analysis
The heterogeneity of the significant results was assessed using Cochran’s Q test and I2 statistics, as shown in Table 3. The results showed moderate heterogeneity (Cochran’s Q p-value < 0.05, 25% > I2 > 50%) among significant anemia indicators in the MR results of AR (IVW). A funnel plot of the instrumental variable of anemia (Fig. 3) showed that the scatter points of the causal association effect were symmetrical, indicating no potential bias in the results. For indicators with highly heterogeneous results in the heterogeneity test, the random-effects model of IVW was used to estimate the causal effects, and the results are shown in Fig. 4 and Table 4.
Table 3
Mendelian randomization analysis heterogeneity test for the association between Anaemia and Allergic Rhinitis
Exposure
|
method
|
Q
|
Q df
|
Q pvalue
|
I2 (%)
|
Other and unspecified anaemias
|
MR Egger
|
16.34993
|
8
|
0.037637
|
0.510701
|
|
Inverse variance weighted
|
17.70284
|
9
|
0.038782
|
0.491607
|
Q, Cochran's Q test statistic;Q df, degrees of freedom for the Q test; I2 statistic reflects the proportion of heterogeneity attributed to instrumental variables in the total variability. |
Table 4
Inverse variance weighted random effects model analysis of the association between anemia and allergic rhinitis
Exposure
|
Number of SNPs
|
Beta
|
Standard error
|
p-value
|
Other and unspecified anemias
|
10
|
0.020549
|
0.010396
|
0.048088
|
SNPs: single nucleotide polymorphisms |
The MR-Egger regression was used to test the horizontal pleiotropy of the instrumental variables. The statistical hypothesis test P-values of the intercept terms of each index exceeded 0.05, and the intercept was close to 0, indicating that the causal inference in this study was not influenced by horizontal pleiotropy (Table 5).
Table 5
Mendelian randomization analysis of anemia’s impact on allergic rhinitis using a horizontal pleiotropy test
Exposure
|
MR-Egger intercept
|
Standard error
|
p-value
|
Other and unspecified anemias
|
0.003878
|
0.004766
|
0.439405
|
Sensitivity analysis of the results was conducted using the elimination test individually (Fig. 5). Each line in the figure represents the effect size and 95% confidence interval range of the index after the corresponding SNP is removed, with the red line representing the reference effect interval. Each line closely overlaps with the red-line interval, indicating that the effect estimate remains stable even with the removal of a single SNP. This suggests that the results are robust.
The results of the reverse causality MR analysis showed no significant causal effect of AR on the anemia index (p-value > 0.05), as shown in Fig. 6.
3.5 Multivariate MR analysis
We conducted a univariate MR analysis of all smoking- and air pollution-related indicators retrieved from the MRC IEU OpenGWAS database and the outcomes of AR, respectively. Significant causal relationships were identified using a combination of IVW models (Table 6). The results indicated significant causal associations among smoking [cigarettes smoked per day (GWAS ID: ieu-b-142), cigarettes smoked per day (GWAS ID: ieu-b-4826)], type of tobacco previously smoked [manufactured cigarettes (GWAS ID: ukb-d-2877_1)], air pollution [particulate matter air pollution (pm10); 2010 (GWAS ID: ukb-e-24005_EAS)], and other indicators. A reverse Mendelian analysis was then performed on the combinations with significant causal relationships in the previously screened IVW model. The results of reverse causality MR analysis showed that AR had a significant effect on smoking [cigarettes smoked per day (GWAS ID: ieu-b-142), cigarettes smoked per day (GWAS ID: ieu-b-4826)], type of tobacco previously smoked [manufactured cigarettes (GWAS ID: ukb-d-2877_1)], and particulate matter air pollution [(pm10), 2010 (GWAS ID: ukb-e-24005_EAS)]. Other indicators had no causal effect (p > 0.05), and the results are shown in Table 7.
Table 6
Causal effect estimates of Mendelian randomization for smoke, air pollution, and allergic rhinitis
Exposure
|
Outcome
|
Number of SNPs
|
Beta
|
Standard error
|
OR95CI
|
p-value
|
Cigarettes smoked per day (GWAS ID: ieu-b-142)
|
Doctor diagnosed hay fever or allergic rhinitis
|
68
|
0.015890
|
0.006535
|
0.984235 (0.971708, 0.996924)
|
0.015043
|
Cigarettes smoked per day (GWAS ID: ieu-b-4826)
|
Doctor diagnosed hay fever or allergic rhinitis
|
19
|
0.003301
|
0.001456
|
0.996704 (0.993863, 0.999552)
|
0.023366
|
Type of tobacco previously smoked: manufactured cigarettes (GWAS ID: ukb-d-2877_1)
|
Doctor diagnosed hay fever or allergic rhinitis
|
16
|
0.168094
|
0.071974
|
0.845274 (0.734061, 0.9733371)
|
0.019518
|
Particulate matter air pollution (pm10); 2010 (GWAS ID: ukb-e-24005_EAS)
|
Doctor diagnosed hay fever or allergic rhinitis
|
6
|
0.014345
|
0.004838
|
1.014448 (1.004874, 1.024113)
|
0.003026
|
SNP, single nucleotide polymorphism; β, the effect coefficients in Mendelian randomization analysis; OR, odds ratio; CI, confidence interval.
Table 7
Reverse Mendelian randomization analysis of smoking (smoke) and air pollution (air pollution) on allergic rhinitis
Outcome
|
Exposure
|
Number of SNPs
|
Beta
|
Standard error
|
p-value
|
Cigarettes smoked per day (GWAS ID: ieu-b-142)
|
Cigarettes smoked per day
|
46
|
0.01349
|
0.086104
|
0.875502
|
Cigarettes smoked per day (GWAS ID: ieu-b-4826)
|
Cigarettes smoked per day
|
37
|
0.00823
|
0.875349
|
0.992498
|
Type of tobacco previously smoked: manufactured cigarettes (GWAS ID: ukb-d-2877_1)
|
Type of tobacco previously smoked: manufactured cigarettes
|
48
|
0.00463
|
0.017885
|
0.795819
|
Particulate matter air pollution (pm10); 2010 (GWAS ID: ukb-e-24005_EAS)
|
Doctor diagnosed hay fever or allergic rhinitis
|
38
|
0.45725
|
0.414422
|
0.269879
|
SNP, single nucleotide polymorphism; β, the effect coefficients in the Mendelian randomization analysis.
Finally, a multivariate MR Analysis of AR was conducted separately using these significant results as exposures to assess the direct effect of anemia on AR (Table 8). Models 1, 2, and 3, adjusted for cigarettes smoked per day (GWAS ID: ieu-b-142), cigarettes smoked per day (GWAS ID: ieu-b-4826), type of tobacco previously smoked, and manufactured cigarettes (GWAS ID: ukb-d-2877_1), revealed no significant effect of anemia on AR. Model 4, adjusted for particulate matter air pollution [(pm10); 2010 (GWAS ID: ukb-e-24005_EAS)], indicated that anemia still exerted a significant direct effect on AR.
Table 8
Results of multivariate Mendelian randomization analysis of the effects of smoking, air pollution, and anemia on the incidence of allergic rhinitis
Model
|
Exposure
|
Outcome
|
Beta
|
Standard error
|
p- value
|
Model 1
|
Other and unspecified anemias (GWAS id: finn-b-D3_ANAEMIANAS)
|
Doctor diagnosed hay fever or allergic rhinitis (GWAS id: ukb-a-254)
|
0.010220
|
0.007825
|
0.191556
|
|
Cigarettes smoked per day (GWAS id: ieu-b-142)
|
Doctor diagnosed hay fever or allergic rhinitis (GWAS id: ukb-a-254)
|
0.018907
|
0.006319
|
0.002771
|
Model 2
|
Other and unspecified anemias (GWAS id: finn-b-D3_ANAEMIANAS)
|
Doctor diagnosed hay fever or allergic rhinitis (GWAS id: ukb-a-254)
|
0.012454
|
0.010209
|
0.222482
|
|
Cigarettes smoked per day (GWAS id: ieu-b-4826)
|
Doctor diagnosed hay fever or allergic rhinitis (GWAS id: ukb-a-254)
|
0.003668
|
0.001390
|
0.008295
|
Model 3
|
Other and unspecified anemias (GWAS id: finn-b-D3_ANAEMIANAS)
|
Doctor diagnosed hay fever or allergic rhinitis (GWAS id: ukb-a-254)
|
0.017658
|
0.009310
|
0.057864
|
|
Type of tobacco previously smoked: manufactured cigarettes (GWAS id: ukb-d-2877_1)
|
Doctor diagnosed hay fever or allergic rhinitis (GWAS id: ukb-a-254)
|
0.120096
|
0.093594
|
0.199436
|
Model 4
|
Other and unspecified anemias (GWAS id: finn-b-D3_ANAEMIANAS)
|
Doctor diagnosed hay fever or allergic rhinitis (GWAS id: ukb-a-254)
|
0.0302924
|
0.013059
|
0.020361
|
|
Particulate matter air pollution (pm10); 2010 (GWAS id: ukb-e-24005_EAS)
|
Doctor diagnosed hay fever or allergic rhinitis (GWAS id: ukb-a-254)
|
0.0137253
|
0.006996
|
0.049762
|
SNP, single nucleotide polymorphism; β, the effect coefficients in the multivariate Mendelian randomization analysis.