Abstract Background Epidemiological associations between maternal concentrations of perfluoroalkyl substances (PFAS) and birth weight are inconsistent. There is concern that studies based on samples collected in late pregnancy may be confounded by kidney function but studies of the relation between pregnancy-induced changes in PFAS and kidney function are lacking. Our aims were to investigate changes in serum concentrations of perfluorononanoic acid (PFNA), perfluorooctanoic acid (PFOA), perfluorooctane sulfonate (PFOS) and perfluorohexane sulfonate (PFHxS) from early to late pregnancy and to explore relations to changes in glomerular filtration rate (GFR) and glomerular pore size. Methods We conducted the study in a cohort of 73 pregnancies of normal-weight Swedish women without gestational diabetes and preeclampsia, enrolled 2009–2014. Blood was collected in median weeks 11 and 36, respectively, and analysed PFAS using liquid chromatography-tandem-mass-spectrometry. We estimated GFR based on creatinine and cystatin C and used the ratio eGFR_{cystatin C}/eGFR_creatinine to indicate glomerular pore size. We used Wilcoxon signed-rank test to compare early and late measures and partial Spearman rank correlations to explore relations between changes in PFAS and kidney function. Results Median concentrations of PFNA, PFOA and PFOS decreased by 15–21% but changes were uncorrelated to changes in kidney function (partial R = − 0.06–0.11). The observed increase in median PFHxS concentration of 69% was likely an artefact of systematic measurement error caused by coeluting endogenous inferences. Conclusions Serum concentrations of PFNA, PFOA and PFOS decrease during pregnancy but the magnitudes of change are unrelated to parallel changes in eGFR and glomerular pore size, suggesting that changes in these indicators of kidney function are not important confounders in studies of PFAS and birth weight in pregnancies without gestational diabetes and preeclampsia.