Table 2.
Incidence rate ratios and differences for leukaemia comparing children prenatally exposed to nitrofurantoin and pivmecillinam
| Any leukaemia |
Lymphoid leukaemia |
AML | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | Cases | IRR (95% CI) | wIRRb (95% CI) | IRD pr 100 000 person-years (95% CI) | wIRDb pr 100 000 person-years (95% CI) | Cases | IRR (95% CI) | wIRRb (95% CI) | IRD pr 100 000 person-years (95% CI) | wIRDb pr 100 000 person-years (95% CI) | Cases | |
| Pivmecillinam | 247 306 | 129 | Reference | 108 | Reference | 13 | ||||||
| Nitrofurantoin | 44 091 | 32 | 1.36 (0.90, 2.06)a | 1.34 (0.88,2.06)a | 1.90 (–1.42, 5.23)a |
1.49 (–1.92, 4.90)a | 26 | 1.42 (0.92, 2.17) | 1.34 (0.83, 2.17) | 1.95 (–0.89, 4.79 | 1.60 (–1.46, 4.66) | <5 |
| Two or more prescription fills | ||||||||||||
| Pivmecillinam | 47 708 | 23 | Reference | 17 | Reference | <5 | ||||||
| Nitrofurantoin | 5643 | 6 | 2.15 (0.88, 5.28) | 1.57 (0.54, 4.55) | 6.24 (–4.24, 16.74) | 3.09 (–5.92, 12.10) | <5 | – | – | – | – | <5 |
| Trimester of exposure | ||||||||||||
| First trimester | ||||||||||||
| Pivmecillinam | 75 160 | 26 | Reference | 24 | Reference | <5 | ||||||
| Nitrofurantoin | 11 688 | 7 | 1.83 (0.79, 4.21) | 1.92 (0.84, 4.41) | 3.11 (–2.61, 8.82) | 3.53 (–2.60, 9.65) | 5 | 1.42 (0.54, 3.71) | 1.48 (0.57, 3.87) | 1.44 (–3.28, 6.15) | 1.71 (–3.35, 6.76) | <5 |
| Second trimester | ||||||||||||
| Pivmecillinam | 88 231 | 48 | Reference | 44 | Reference | 0 | ||||||
| Nitrofurantoin | 15 657 | 6 | 0.74 (0.32, 1.72) | 0.53 (0.19, 1.47) | –1.58 (–5.31, 2.16) | –2.77 (–5.97, 0.43) | 5 | 0.67 (0.27, 1.69) | 0.58 (0.21, 1.61) | –1.81 (–5.20, 1.58) | –2.28 (–5.49, 0.93) | <5 |
| Third trimester | ||||||||||||
| Pivmecillinam | 116 817 | 67 | Reference | 51 | Reference | 11 | ||||||
| Nitrofurantoin | 20 460 | 21 | 1.73 (1.03, 2.92) | 1.73 (1.00, 2.98) | 4.81 (–1.13, 10.76) | 4.77 (–1.37, 10.90) | 16 | 1.63 (0.89, 3.00) | 1.66 (0.88, 3.11) | 3.31 (–1.90, 8.52) | 3.44 (–2.01, 8.88) | <5 |
Results from fixed-effects meta-analysis, I2 = 0.0%. Findings from mixed-effects Poisson models were comparable: IRR 1.44 (0.96, 2.17), wIRR 1.40 (0.91, 2.15).
Inverse probability of treatment weights including calendar year at birth, maternal age, parity, maternal history of cancer before pregnancy, prescription fills for immunosuppressants, systemic corticosteroids and systemic antibiotics before start of pregnancy, maternal smoking status during first trimester and child sex. In Finland, birth year was not balanced after weighting and hence was added to the outcome model.
AML, acute myeloid leukaemia; IRD, incidence rate difference; IRR, incidence rate ratio; wIRD, weighted incidence rate difference; wIRR, weighted incidence rate ratio.