We studied 2040 women aged 14 to 48, with a mean age of 26.56 years old. Most women (93.69%) were on their first (19.46%), second (18.58%), or third (7.84%) pregnancies. Among all participants, 464 (22.75%) were from rural areas, and the rest (77.25%) were urbanites. The proportion of congenital malformation was low (0.78%), and the records showed that only 42 (2.06%) of neonates experienced NCs. Detailed information about the maternal demographic information, prepartum events, NCs, and type of delivery is available in Appendix.
In total number of mothers who dyed their hair during pregnancy (75.24%), the majority (62.67%) colored their hair in the third trimester. There was no statistically significant association between hair coloring and neonate 1-minute (p = 0.23) and 5-minute Apgar score (p = 0.99). In subgroup analysis, Chi-square test of independence with post hoc tests showed that the neonates of the women who dyed their hair in the preconception period had more unfavorable 1-minute Apgar scores (p < 0.0001)
Chi-square test of independence results of comparison for overall NCs and birth weight categories according to hair coloring indices are displayed in Table 1 and Table 2.
Table 1
Comparisons of the Frequency of Neonatal Complications According to Coloring Characteristics
Characteristics | Overall sample | Complication | No complication | Chi-square tests of independence |
Period of coloring | | | | |
No coloring | 505(24.75%) | 22(52.38%) | 483(24.17%) | ꭓ2(4) = 85.10 p < 0.001 n = 2040 |
Pre-pregnancy | 159(7.79%) | 16 (38.10%) | 143(7.16%) |
First trimester | 343(16.81%) | 1(2.38%) | 342 (17.12%) |
Second trimester | 71(3.48%) | 0(0.00%) | 71(3.55%) |
Third trimester | 962(47.16%) | 3(7.14%) | 959(48.00%) |
Product of color | | | | |
No Coloring | 505(24.75%) | 22(52.38%) | 483(24.17%) | ꭓ2(3) = 18.73 p < 0.001 n = 2040 |
Domestic product | 1023(50.15%) | 14(33.33%) | 1009(50.50%) |
Foreign Product | 311(15.25%) | 2(4.76%) | 309(15.47%) |
Unknown | 201(9.85%) | 4(9.52%) | 197(9.86%) |
Numbers of standard color tube used | | | |
No Coloring | 505(24.75%) | 22(52.38%) | 483(24.17%) | ꭓ2(4) = 33.67 p < 0.001 n = 2040 |
0. 5 Tube | 362(17.75%) | 6(14.29%) | 356(17.82%) |
One Tube | 973(47.70%) | 5(11.90%) | 986(48.45%) |
Two Tubes | 176(8.63%) | 9(21.43%) | 167(8.36%) |
Four Tubes | 24(1.18%) | 0.0(%) | 24(1.20%) |
Style Of coloring | | | | |
No Coloring | 505(24.75%) | 22(52.38%) | 483(24.17%) | ꭓ2(4) = 27.96 p < 0.001 n = 2040 |
Bleaching | 36(1.76%) | 3(7.14%) | 33(1.65%) |
Simple coloring | 1180(57.84%) | 16(38.10%) | 1164(58.26%) |
Salt and pepper | 319(15.64%) | 1(2.38%) | 318(15.92%) |
Color of dye | | | | |
Bright | 428(27.88%) | 1(5.00%) | 427(28.18%) | ꭓ2(1) = 5.27 p = 0.02 n = 1535 |
Dark | 1107 (72.12%) | 19(95.00%) | 1088(71.82%) |
Dark Or Bright | 0(0%) | 0(0%) | 0(0%) |
=effect size (phi coefficient or Cramer’s V). a: A significant difference with an effect size ≥ Cohen’s definition of “Medium”.b: A significant difference with an effect size ≥ Cohen’s definition of “small.” |
Table 2
Comparisons of the Newborn’s Weight Categories According to Coloring Characteristics
Characteristics | Overall sample | VLBW | LBW | NBW* | HBW | Chi-square tests of independence |
Period of coloring | | | | | | |
No coloring | 505(24.75%) | 1(11.11%) | 35(23.18%) | 447(24.53%) | 22 (37.93%) | ꭓ2(12) = 77.39 p < 0.001 n = 2040 |
Pre-pregnancy | 159(7.79%) | 5(55.56%) | 24(15.89%) | 121(6.64%) | 9(15.52%) |
First trimester | 343(16.81%) | 3(33.33%) | 38(25.17%) | 296 (16.25%) | 6(10.34%) |
Second trimester | 71(3.48%) | 0(0.00%) | 5(3.31%) | 66 (3.62%) | 0(0.00%) |
Third trimester | 962(47.16%) | 0(0.00%) | 49(32.45%) | 892 (48.96%) | 21(36.21%) |
Product of dye | | | | | | |
No Coloring | 505(24.75%) | 1(11.11%) | 35(23.18%) | 447(24.53%) | 22(37.93%) | ꭓ2(9) = 18.89 p = 0.026 n = 2040 |
Domestic product | 1023(50.15%) | 5(55.56%) | 73(48.34%) | 926(50.82%) | 19(32.76%) |
Foreign Product | 311(15.25%) | 0(0.00%) | 29(19.21%) | 274(15.04%) | 8(13.79%) |
Unknown | 201(9.85%) | 3(33.33%) | 14(9.27%) | 175(9.60%) | 9(15.52%) |
Amount of standard color tube used | | | | |
No Coloring | 505(24.75%) | 1(11.11%) | 35(23.18%) | 447(24.53%) | 22(37.93%) | ꭓ2(12) = 24.86 p = 0.016 n = 2040 |
0. 5 Tube | 362(17.75%) | 1(11.11%) | 21(13.91%) | 328(18.00%) | 12(20.69%) |
One Tube | 973(47.70%) | 7(77.78%) | 68(45.03%) | 880(48.30%) | 18(31.03%) |
Two Tubes | 176(8.63%) | 0(0.00%) | 24(15.89%) | 146(8.01%) | 6(8.63%) |
Four Tubes | 24(1.18%) | 0(0.00%) | 3(1.99%) | 21(1.15%) | 0(0.00%) |
Color of dye | | | | | | |
Bright | 428(27.88%) | 4(50.00%) | 24(20.69%) | 393(28.58%) | 7(19.44%) | ꭓ2(3) = 6.54 p = 0.088 n = 2040 |
Dark | 1107(72.12%) | 4(50.00%) | 92(79.31%) | 982(71.42%) | 29(80.56%) |
Dark or Bright | 0(0%) | 0(0%) | 0(0%) | 0(0%) | 0(0%) |
Hair Coloring Style |
No Coloring | 505(24.75%) | 1(11.11%) | 35(23.18%) | 447(24.53%) | 22(37.93%) | ꭓ2(9) = 46.28 p < 0.001 n = 2040 |
Bleaching | 36(1.76%) | 2(22.22%) | 3(1.99%) | 28(1.54%) | 3(5.17%) |
Simple coloring | 1180(57.84%) | 5(55.56%) | 104(68.87%) | 1046(57.41%) | 25(43.10%) |
Salt and pepper | 319(15.64%) | 1(11.11%) | 9(5.96%) | 301(16.52%) | 8(13.79%) |
=effect size (phi coefficient or Cramer’s V). a: A significant difference with an effect size ≥ Cohen’s definition of “small”.b: not statistically significant. NBW (normal birth weight) |
NCs were more prevalent in the women, who had never dyed their hair before conception, than in others. (p < 0.001). Therefore, we ran a subgroup analysis with post hoc chi-square test to examine the differences in NCs in the participants who dyed their hair preconception or during pregnancy. Test results revealed that the recorded NCs were higher in women who dyed their hair in the preconception period in comparison to pregnancy trimesters (p < 0.001)1. Similarly, the post-hoc chi-square test showed that using approximately two tubes of coloring material was significantly associated with higher rates of neonatal adverse outcomes (p = 0.002).
When comparing NCs in the categories of “dye production” and “style of hair coloring”, NCs were recorded more frequently in mothers who did not color their hair (p < 0.001). However, the prevalence of NCs was statistically similar in different categories of dye product and the style of hair coloring (p > 0.006).
We performed logistic regression analysis to ascertain the effects of the period of using hair color, the product of the dye, the amount of the hair dye used, and the style of hair coloring on the likelihood that neonates experience adverse outcomes. The logistic regression model was statistically significant (χ2 (11) = 82.71, p < .001). The model explained 21.9% (Nagelkerke R2) of the variance in neonatal adverse outcomes and correctly classified 98.1% of cases. Neonates born to women who dyed their hair in the preconception period were less likely to exhibit NCs than babies of the women who did not use hair dye and the women who dyed their hair during pregnancy trimesters (OR:0.027 CI95%:0.007–0.108, p < 0.001). The likelihood of adverse neonate outcomes was not associated with other hair coloring variables (p > 0.1).
A Chi-Square test of independence revealed that birth weight categories were significantly associated with the hair coloring period (p < 0.001). Post-hoc comparisons of rates of LBW by the period of hair coloring revealed that higher rates of LBW were seen among those dyed their hair in the preconception period and third trimester (p < 0.001). The LBW prevalence was statistically similar among those who colored their hair in the first and second trimester and those who did not color their hair (p > 0.002). Very low birth weight (VLBW) and high birth weight (HBW) prevalence were statistically similar among all groups (p > 0.002).
Although the Chi-Square test of independence showed significant associations between birth weight categories and the production of the dye, hair coloring style, and the amount of color used, post-hoc tests exposed that rates of VLBW, LBW, and HBW were statistically similar among all groups (Table 2).
Chi-square test of independence depicted a higher prevalence of the abruptio placenta (p < 0.001) and PROM (p < 0.001) in women who did not dye their hair. On the other hand, preterm labor recorded at higher rates in the women who used hair coloring (ꭓ2(4) = 43.94,p < 0.001,). Subgroup comparison with post hoc chi-square test revealed that women who dyed their hair on the third-trimester experienced statistically significant higher rates of preterm labor (p < 0.0001).