Table 3.

Adjusted ORsa for IVF/ICSI failures in relation to serum PCB quartiles (Q) among cycles with an embryo transfer.b

Table 3. Adjusted ORsa for IVF/ICSI failures in relation to serum PCB quartiles (Q) among cycles with an embryo transfer.b
		No. of cycles		No. of women		Failed implantation				Chemical pregnancy				Spontaneous abortion
Exposure quartiles						No. of events		ORs (95% CI)		No. of events		ORs (95% CI)		No. of events		ORs (95% CI)
PCB-118
Q1		196		185		43		Referent		54		Referent		25		Referent
Q2		196		183		59		1.81 (1.11–2.97)		38		0.72 (0.43–1.22)		34		1.50 (0.79–2.83)
Q3		190		176		71		2.34 (1.42–3.84)		38		0.89 (0.52–1.51)		33		1.70 (0.87–3.32)
Q4		192		185		61		1.60 (0.96–2.69)		47		0.91 (0.54–1.53)		41		1.29 (0.66–2.53)
p for trend								0.06				0.87				0.42
PCB-138
Q1		197		184		49		Referent		45		Referent		31		Referent
Q2		198		189		59		1.42 (0.87–2.30)		41		0.98 (0.58–1.68)		25		0.68 (0.36–1.34)
Q3		185		170		62		1.51 (0.91–2.48)		46		1.18 (0.68–2.05)		30		1.10 (0.57–2.15)
Q4		194		190		64		1.39 (0.83–2.30)		45		1.10 (0.63–1.91)		38		1.14 (0.59–2.20)
p for trend								0.24				0.61				0.41
PCB-153
Q1		197		188		46		Referent		45		Referent		30		Referent
Q2		193		182		57		1.63 (0.99–2.71)		42		1.04 (0.60–1.78)		26		0.89 (0.46–1.73)
Q3		192		179		59		1.59 (0.95–2.66)		46		1.09 (0.63–1.90)		33		1.18 (0.61–2.27)
Q4		192		186		72		1.99 (1.16–3.40)		44		1.17 (0.66–2.10)		35		1.13 (0.56–2.27)
p for trend								0.02				0.58				0.56
Group 1 PCBs
Q1		195		186		59		Referent		41		Referent		28		Referent
Q2		194		187		50		0.87 (0.54–1.41)		46		1.06 (0.63–1.80)		34		1.20 (0.64–2.26)
Q3		195		187		57		0.99 (0.61–1.61)		44		0.97 (0.56–1.68)		30		0.83 (0.43–1.62)
Q4		190		183		68		1.19 (0.72–1.93)		46		1.13 (0.65–1.98)		32		0.93 (0.46–1.85)
p for trend								0.38				0.76				0.59
Group 2 PCBs
Q1		196		185		51		Referent		46		Referent		27		Referent
Q2		197		187		56		1.27 (0.78–2.08)		41		0.91 (0.53–1.55)		27		0.91 (0.47–1.76)
Q3		191		178		60		1.44 (0.87–2.36)		41		0.97 (0.56–1.68)		39		1.66 (0.86–3.18)
Q4		190		184		67		1.38 (0.82–2.32)		49		1.17 (0.67–2.05)		31		1.07 (0.53–2.14)
p for trend								0.20				0.54				0.44
Group 3 PCBs
Q1		198		189		46		Referent		45		Referent		30		Referent
Q2		191		180		57		1.44 (0.88–2.38)		39		0.90 (0.53–1.55)		30		1.01 (0.53–1.93)
Q3		193		178		59		1.42 (0.85–2.38)		51		1.26 (0.74–2.17)		27		0.94 (0.48–1.85)
Q4		192		187		71		1.75 (1.03–2.97)		42		1.02 (0.57–1.84)		37		1.13 (0.57–2.25)
p for trend								0.05				0.65				0.79
∑PCB
Q1		197		187		46		Referent		42		Referent		28		Referent
Q2		194		186		52		1.13 (0.68–1.85)		43		1.08 (0.69–1.85)		33		1.19 (0.63–2.26)
Q3		193		182		58		1.33 (0.81–2.18)		48		1.34 (0.78–2.31)		32		1.36 (0.70–2.65)
Q4		190		184		73		1.70 (1.02–2.85)		44		1.27 (0.71–2.29)		31		1.07 (0.53–2.15)
p for trend								0.03				0.30				0.75
aJoint models for multiple outcomes and multiple cycles per woman were used. Models were adjusted for serum lipids, study site, study phase, race/ethnicity, previous live birth, maternal age, BMI, smoking status, ampules of gonadotropins, protocol, ICSI, number of embryos transferred, and primary infertility diagnosis. bThere were 827 cycles with serum samples analyzed for PCB concentrations. However, because of missing covariates, only 774 cycles from 720 women contributed to the multivariate analysis. Missing ethnicity, infertility diagnosis, and previous live birth were recoded as a separate category so that cycles missing this information were in the multivariate models. If a cycle was missing any other covariate, that cycle was dropped from the multivariate analysis. Specifically, 13 cycles were missing lipids, 2 cycles were missing ampules of gonadotropins, 32 cycles were missing number of embryos transferred, and 6 cycles were missing BMI.