Does the growing of Bt maize change abundance or ecological function of non-target animals compared to the growing of non-GM maize? A systematic review

In some analyses, a positive effect of Bt maize was observed for flower bugs (Anthocoridae).

“and a positive effect for ladybeetles (Coccinellidae), flower bugs (Anthocoridae), and lacewings (Neuroptera).”

The database on non-target effects of Bt maize field trials contains more than 7200 records from 233 experiments and 120 articles.

“The database on non-target effects of Bt maize field trials contains more than 7200 records from 233 experiments and 120 articles.”

Few data were available for insecticides other than pyrethroids.

“In particular, populations of predators were reduced after pyrethroid treatment, while few data were available for other insecticides.”

Syrphidae showed a negative effect size with an upper boundary of exact zero.

“Syrphidae showed a negative effect size with an upper boundary of exact zero.”

• AssumptionThe authors state this as an assumption: For plot and split-field designs, the history of management was equal between Bt and non-Bt plots because plots were installed within the same field.
• AssumptionThe authors state this as an assumption: For meta-analyses, when authors did not mention pesticide treatments or when Bt and non-Bt plots received similar treatments, no pesticides were applied.
• AssumptionThe authors state this as an assumption: If commercial seeds were used, sufficient Bt protein expression was assumed.
• AssumptionThe authors state this as an assumption: If no information on insecticides before the start of the study was provided, no insecticides were applied.

Significant heterogeneity was observed for Lathrididae (Coleoptera) and Otitidae (Diptera).

“Significant heterogeneity was observed for Lathrididae (Coleoptera) and Otitidae (Diptera), but both groups were represented by records from only 4 and 3 articles (8 and 21 experiments), respectively.”

• AssumptionThe authors state this as an assumption: For plot and split-field designs, the history of management was equal between Bt and non-Bt plots because plots were installed within the same field.
• AssumptionThe authors state this as an assumption: For meta-analyses, when authors did not mention pesticide treatments or when Bt and non-Bt plots received similar treatments, no pesticides were applied.
• AssumptionThe authors state this as an assumption: If commercial seeds were used, sufficient Bt protein expression was assumed.
• AssumptionThe authors state this as an assumption: If no information on insecticides before the start of the study was provided, no insecticides were applied.

Meta-analyses on different taxonomic levels revealed only few and often non-robust significant effect sizes when both Bt maize and non-Bt maize were untreated.

“Meta-analyses on different taxonomic levels revealed only few and often non-robust significant effect sizes when both Bt maize and non-Bt maize were untreated.”

In some analyses, a positive effect of Bt maize was observed for lacewings (Neuroptera).

“and a positive effect for ladybeetles (Coccinellidae), flower bugs (Anthocoridae), and lacewings (Neuroptera).”

Bt maize harboured fewer parasitoids (Braconidae, Tachinidae) of the European corn borer compared with non-Bt maize.

“Bt maize harboured fewer parasitoids (Braconidae, Tachinidae) of the European corn borer, the main target pest of Lepidoptera-active Bt maize, compared with non-Bt maize.”

Meta-analyses with all Bt proteins and excluding red-flagged records showed reduced abundance, activity density, or predation/parasitisation rates in untreated Bt maize compared with untreated non-Bt maize for Staphylinidae (Coleoptera), Tachinidae (Diptera), and Braconidae (Hymenoptera).

“Meta-analyses with all Bt proteins, but without red flagged records in the critical appraisal revealed reduced abundance, activity density, or predation/parasitisation rates in untreated Bt maize compared with untreated non-Bt maize for Staphylinidae (Coleoptera), Tachinidae (Diptera), and Braconidae (Hymenoptera) (Fig. 8, Additional file 7: Table S7.1).”

• AssumptionThe authors state this as an assumption: For plot and split-field designs, the history of management was equal between Bt and non-Bt plots because plots were installed within the same field.
• AssumptionThe authors state this as an assumption: For meta-analyses, when authors did not mention pesticide treatments or when Bt and non-Bt plots received similar treatments, no pesticides were applied.
• AssumptionThe authors state this as an assumption: If commercial seeds were used, sufficient Bt protein expression was assumed.
• AssumptionThe authors state this as an assumption: If no information on insecticides before the start of the study was provided, no insecticides were applied.

In some analyses, a negative effect of Bt maize was observed for rove beetles (Staphylinidae).

“In some analyses, a negative effect of Bt maize was observed for rove beetles (Staphylinidae) and hoverflies (Syrphidae)”

In some analyses, a positive effect of Bt maize was observed for ladybeetles (Coccinellidae).

“and a positive effect for ladybeetles (Coccinellidae), flower bugs (Anthocoridae), and lacewings (Neuroptera).”

Populations of predators were reduced after pyrethroid treatment.

“In particular, populations of predators were reduced after pyrethroid treatment, while few data were available for other insecticides.”

Effect sizes for order-level analyses in Diptera and Hymenoptera were significant.

“Effect sizes for order-level analyses in Diptera and Hymenoptera also were significant”

• AssumptionThe authors state this as an assumption: For plot and split-field designs, the history of management was equal between Bt and non-Bt plots because plots were installed within the same field.
• AssumptionThe authors state this as an assumption: For meta-analyses, when authors did not mention pesticide treatments or when Bt and non-Bt plots received similar treatments, no pesticides were applied.
• AssumptionThe authors state this as an assumption: If commercial seeds were used, sufficient Bt protein expression was assumed.
• AssumptionThe authors state this as an assumption: If no information on insecticides before the start of the study was provided, no insecticides were applied.

In some analyses, a negative effect of Bt maize was observed for hoverflies (Syrphidae).

“In some analyses, a negative effect of Bt maize was observed for rove beetles (Staphylinidae) and hoverflies (Syrphidae)”

Funnel plots showed no evidence for publication bias.

“Funnel plots showed no evidence for publication bias and the analyses of private sector contribution revealed no evidence for influence of vested interests.”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, more effect sizes were significant.

“When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, more effect sizes were significant.”

Sap beetles (Nitidulidae), that are associated with Lepidoptera damage, were recorded less in Bt maize.

“Similarly, sap beetles (Nitidulidae), that are associated with Lepidoptera damage, were recorded less in Bt maize.”

Hymenoptera showed significant heterogeneity.

“and Hymenoptera showed significant heterogeneity.”

• AssumptionThe authors state this as an assumption: For plot and split-field designs, the history of management was equal between Bt and non-Bt plots because plots were installed within the same field.
• AssumptionThe authors state this as an assumption: For meta-analyses, when authors did not mention pesticide treatments or when Bt and non-Bt plots received similar treatments, no pesticides were applied.
• AssumptionThe authors state this as an assumption: If commercial seeds were used, sufficient Bt protein expression was assumed.
• AssumptionThe authors state this as an assumption: If no information on insecticides before the start of the study was provided, no insecticides were applied.

Braconidae revealed significant data heterogeneity.

“Braconidae also revealed significant data heterogeneity.”

• AssumptionThe authors state this as an assumption: For plot and split-field designs, the history of management was equal between Bt and non-Bt plots because plots were installed within the same field.
• AssumptionThe authors state this as an assumption: For meta-analyses, when authors did not mention pesticide treatments or when Bt and non-Bt plots received similar treatments, no pesticides were applied.
• AssumptionThe authors state this as an assumption: If commercial seeds were used, sufficient Bt protein expression was assumed.
• AssumptionThe authors state this as an assumption: If no information on insecticides before the start of the study was provided, no insecticides were applied.

Analyses of private sector contribution revealed no evidence for influence of vested interests.

“Funnel plots showed no evidence for publication bias and the analyses of private sector contribution revealed no evidence for influence of vested interests.”

The total number of records in the database labelled with 'red' in at least one critical appraisal criterion was 911 of 7279 records (12.5%).

“In summary, the total number of records in the database labelled with “red” in at least one critical appraisal criterion, thus indicating low internal or external validity, was 911 of 7279 records (12.5%).”

384 of 1976 records from 5 articles (19.6%) represent experiments where control plot data were used twice (332 records) or four times (52 records) because two or four Bt maize cultivars shared the same control cultivar.

“Altogether, 384 of the 1976 records from 5 articles (19.6%) represent experiments where the data from the control plots were used twice (332 records) or four times (52 records) because two or four Bt maize cultivars had the same control cultivar.”

In analyses restricted to Lepidoptera-active Bt proteins, Bt plots or fields had lower populations of Nitidulidae (significant negative effect size).

“When the analyses were repeated for only Lepidoptera-active Bt proteins, lower populations in Bt plots or fields (significant negative effect size) were observed for Nitidulidae, Tachinidae, and Braconidae”

On higher taxonomic level, Diptera showed negative effect sizes in analyses restricted to Lepidoptera-active Bt proteins.

“On higher taxonomic level, Diptera and Hymenoptera showed negative effect sizes.”

The main analyses with 1976 selected records labelled as comparing 'untreated' Bt and non-Bt maize included 82 records (4%) from studies where Bt and non-Bt plots were treated with the same insecticide.

“The main analyses with 1976 selected records labelled as comparing “untreated” Bt and non-Bt maize included 82 records (4%) from studies where Bt and non-Bt plots were treated with the same insecticide”

In analyses restricted to Lepidoptera-active Bt proteins, Bt plots or fields had lower populations of Tachinidae (significant negative effect size).

“When the analyses were repeated for only Lepidoptera-active Bt proteins, lower populations in Bt plots or fields (significant negative effect size) were observed for Nitidulidae, Tachinidae, and Braconidae”

Only 7 records for pollinators were available (0.1%, 3 experiments).

“Only 7 records for pollinators were available (0.1%, 3 experiments).”

The remaining 30% of the records, labelled 'yellow', were based on three or fewer samplings.

“The remaining 30% of the records, labelled “yellow”, were based on ≤ three samplings.”

Most red labels were assigned because of low numbers of collected individuals.

“Most red labels were assigned because of low numbers of collected individuals.”

Expression of the Bt protein was often not confirmed by the authors in the field experiments where invertebrates were collected.

“Regarding external validity of a record, which means that the data are relevant for the review question, expression of the Bt protein was often not confirmed by the authors in the field experiments where invertebrates were collected (EX).”

In analyses restricted to Lepidoptera-active Bt proteins, Bt plots had higher populations of Anthocoridae (positive effect size).

“while higher populations in Bt plots were observed for Anthocoridae (positive effect size)”

In 9% of the records less than 20 individuals were collected over the whole season in all plots.

“In 9% of the records less than 20 individuals were collected over the whole season in all plots (LN).”

In 61% of the records in the database (215 experiments), invertebrates recorded in Bt maize without any insecticide treatment were compared with those recorded in non-Bt maize without any treatment.

“In 61% of the records in the database (215 experiments), invertebrates recorded in Bt maize without any insecticide treatment were compared with those recorded in non-Bt maize without any treatment (Fig. 5E, black bar left).”

For 25% of records, no information on pesticide treatments during the study was provided.

“For 25% of records, however, no information on pesticide treatments during the study was provided.”

Less than 450 records from 38 experiments were available from South America, Asia, and Africa combined.

“Less than 450 records from 38 experiments were available from South America, Asia, and Africa combined.”

Data on species level are available for 280 species in 2466 records and 157 experiments.

“Data on species level are available for 280 species in 2466 records and 157 experiments.”

Data heterogeneity was present for Syrphidae in analyses restricted to Lepidoptera-active Bt proteins.

“Data heterogeneity was present for Diptera, Syrphidae, Hymenoptera, and Braconidae.”

On higher taxonomic level, Hymenoptera showed negative effect sizes in analyses restricted to Lepidoptera-active Bt proteins.

“On higher taxonomic level, Diptera and Hymenoptera showed negative effect sizes.”

In analyses restricted to Lepidoptera-active Bt proteins, Bt plots or fields had lower populations of Braconidae (significant negative effect size).

“When the analyses were repeated for only Lepidoptera-active Bt proteins, lower populations in Bt plots or fields (significant negative effect size) were observed for Nitidulidae, Tachinidae, and Braconidae”

Data heterogeneity was present for Diptera in analyses restricted to Lepidoptera-active Bt proteins.

“Data heterogeneity was present for Diptera, Syrphidae, Hymenoptera, and Braconidae.”

SEs were sometimes based on the number of traps or sampled plants rather than the number of true replicates (plots or fields).

“SEs were sometimes based on the number of traps or sampled plants rather than the number of true replicates (plots or fields) (VA).”

In Lepidoptera-active Bt maize, more eggs of Neuroptera were recorded compared with non-Bt maize.

“In Bt maize (Lepidoptera-active), fewer larvae and pupae of Tachinidae and more eggs of Neuroptera were recorded compared with non-Bt maize”

In Lepidoptera-active Bt maize, fewer larvae and pupae of Tachinidae were recorded compared with non-Bt maize.

“In Bt maize (Lepidoptera-active), fewer larvae and pupae of Tachinidae and more eggs of Neuroptera were recorded compared with non-Bt maize”

Predatory species were more abundant in untreated Bt maize compared with pyrethroid-treated non-Bt maize.

“Furthermore, predatory species were more abundant in untreated Bt maize compared with pyrethroid-treated non-Bt maize.”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, positive effect sizes were observed for Cantharidae.

“When untreated Bt maize was compared with pyrethroid treated non-Bt maize (mainly foliar and soil application), positive effect sizes (higher populations in Bt maize) were observed for Araneae, Cantharidae, Coccinellidae, Anthocoridae, and Cicadellidae and on higher taxonomic levels for Coleoptera, Hemiptera, and all taxa”

Inclusion of red-flagged records resulted in a similar outcome with a negative effect size for Tachinidae.

“Inclusion of the red-flagged records resulted in a similar outcome of the analyses with a negative effect size (lower populations in Bt maize) for Staphylinidae, Syrphidae, Tachinidae, and Braconidae, and consequently for the higher taxonomic orders Diptera and Hymenoptera.”

Leave-one-out analyses revealed that the majority of significant effects were robust when one record or one experiment was removed at a time.

“The “leave one out” analyses revealed that the majority of significant effects were robust when one record or one experiment was removed at the time.”

Moderator analyses showed that negative effects on Tachinidae and Diptera were mainly associated with effects of Cry1Ab-producing maize.

“Moderator analyses, which examined the effects of different Bt proteins within a target group, revealed that the negative effects on Tachinidae (and Diptera) and on Braconidae (and Hymenoptera) were mainly associated with effects of Cry1Ab-producing maize.”

For 9 of 39 significant effects, at least one record changed the respective analysis to non-significant when removed.

“For 9 of 39 significant effects, however, at least one record was identified that changed the respective analysis to non-signicficant when removed”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, positive effect sizes were observed for Hemiptera.

“When untreated Bt maize was compared with pyrethroid treated non-Bt maize (mainly foliar and soil application), positive effect sizes (higher populations in Bt maize) were observed for Araneae, Cantharidae, Coccinellidae, Anthocoridae, and Cicadellidae and on higher taxonomic levels for Coleoptera, Hemiptera, and all taxa”

Of the analysed species and subgroups, lower numbers of M. cingulum were recorded in Bt maize compared with non-Bt maize.

“Of those, only 2 showed significant effect sizes: higher numbers of H. axyridis and lower numbers of M. cingulum were recorded in Bt maize compared with non-Bt maize”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, positive effect sizes (higher populations in Bt maize) were observed for Araneae.

“When untreated Bt maize was compared with pyrethroid treated non-Bt maize (mainly foliar and soil application), positive effect sizes (higher populations in Bt maize) were observed for Araneae, Cantharidae, Coccinellidae, Anthocoridae, and Cicadellidae and on higher taxonomic levels for Coleoptera, Hemiptera, and all taxa”

For seed treatments with chloro-nicotinyl insecticides, positive effect sizes were observed for Hemiptera.

“Positive effect sizes were observed for Araneae, Hemiptera and the functional group of predators; no heterogeneity was detected for any analysed taxon or functional group.”

Analyses with only green-flagged records confirmed negative effect sizes for Diptera.

“The negative effect sizes for Diptera, Syrphidae, and Hymenoptera were confirmed.”

For seed treatments with chloro-nicotinyl insecticides, no heterogeneity was detected for any analysed taxon or functional group.

“Positive effect sizes were observed for Araneae, Hemiptera and the functional group of predators; no heterogeneity was detected for any analysed taxon or functional group.”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, negative effect sizes (lower populations in Bt maize) were observed for Nitidulidae.

“In contrast, negative effect sizes (lower populations in Bt maize) were observed for Nitidulidae, Aphididae, and Formicidae.”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, negative effect sizes were observed for Formicidae.

“In contrast, negative effect sizes (lower populations in Bt maize) were observed for Nitidulidae, Aphididae, and Formicidae.”

Analyses with only green-flagged records revealed a positive effect on Coccinellidae.

“In addition, a positive effect on Coccinellidae became apparent.”

Meta-regressions showed no effects of number of years of continuous Bt maize cultivation for any functional group.

“Similarly, the number of years of continuous Bt maize cultivation was analysed on functional group level by meta-regression (Table 1). No effects were evident for any functional group.”

Lower populations in Bt maize were detected with sticky traps for Staphylinidae.

“Lower populations in Bt maize (negative effect sizes) were detected with sticky traps for Staphylinidae, Syrphidae, and Braconidae, and with beat cloths for Aphididae.”

Stacked Lepidoptera- and Coleoptera-active maize revealed a negative effect size for Diptera.

“Stacked Lepidoptera and Coleoptera-active maize revealed a negative effect size for Diptera”

Moderator analyses showed that negative effects on Braconidae and Hymenoptera were mainly associated with effects of Cry1Ab-producing maize.

“Moderator analyses, which examined the effects of different Bt proteins within a target group, revealed that the negative effects on Tachinidae (and Diptera) and on Braconidae (and Hymenoptera) were mainly associated with effects of Cry1Ab-producing maize.”

An additional analysis with Lepidoptera-active maize and parasitoids, excluding parasitoids of target O. nubilalis larvae, based on 71 records from 43 experiments and 22 articles, had an effect size of 0.0032 [−0.1704; 0.1640] without heterogeneity (Q = 44.74; p = 0.99).

“Based on 71 records from 43 experiments and 22 articles, the effect size was 0.0032 [− 0.1704; 0.1640] without heterogeneity (Q = 44.74; p = 0.99).”

Analyses on functional group level revealed no effect of Bt maize on predators.

“Analyses on functional group level revealed no effect of Bt maize on decomposers, omnivores, and predators”

Moderator analyses revealed negative effect sizes in Lepidoptera-active Bt maize for Syrphidae collected with sticky traps.

“Moderator analyses revealed negative effect sizes in Lepidoptera-active Bt maize for Syrphidae collected with sticky traps and in Coleoptera-active Bt maize for Chrysomelidae collected with sticky traps”

Moderator analysis for parasitoids was significant for Cry1Ab.

“Parasitoids showed lower populations in Bt maize (negative effect size) in the analysis with all Bt proteins (and significant heterogeneity), and in the analysis with Lepidoptera-active Bt maize (significant heterogeneity; moderator analysis significant for Cry1Ab).”

When the private sector contributed to the data and all Bt proteins were analysed together, lower invertebrate populations in Bt maize (negative effect size) were observed.

“When the private sector contributed to the data and all Bt proteins were analysed together, lower invertebrate populations in Bt maize (negative effect size) were observed.”

Moderator analyses revealed negative effect sizes in Coleoptera-active Bt maize for Chrysomelidae collected with sticky traps.

“Moderator analyses revealed negative effect sizes in Lepidoptera-active Bt maize for Syrphidae collected with sticky traps and in Coleoptera-active Bt maize for Chrysomelidae collected with sticky traps”

Analyses with only green-flagged records confirmed negative effect sizes for Syrphidae.

“The negative effect sizes for Diptera, Syrphidae, and Hymenoptera were confirmed.”

15 significant effects turned non-significant when at least one experiment was removed.

“and 15 significant effects turned to non-significant when at least one experiment was removed.”

More predators were collected in untreated Bt maize compared to non-Bt maize treated with organophosphorous or microbial insecticides.

“In contrast, more predators were collected in untreated Bt maize compared to non-Bt maize treated with organophosphorous or microbial insecticides.”

Positive effect sizes were present in Lepidoptera-active Bt maize for Coccinellidae collected with sticky traps.

“Positive effect sizes were present in Lepidoptera-active Bt maize for Coccinellidae collected with sticky traps and Anthocoridae recorded with visual counts and in stacked Bt maize for Neuroptera recorded with visual counts.”

Inclusion of red-flagged records resulted in a similar outcome with a negative effect size for Syrphidae.

“Inclusion of the red-flagged records resulted in a similar outcome of the analyses with a negative effect size (lower populations in Bt maize) for Staphylinidae, Syrphidae, Tachinidae, and Braconidae, and consequently for the higher taxonomic orders Diptera and Hymenoptera.”

Analysis for herbivores indicated lower populations in stacked Coleoptera- and Lepidoptera-active maize.

“In addition, analysis for herbivores indicated lower populations in stacked Coleoptera- and Lepidoptera-active maize.”

Data heterogeneity was present for Hymenoptera in analyses restricted to Lepidoptera-active Bt proteins.

“Data heterogeneity was present for Diptera, Syrphidae, Hymenoptera, and Braconidae.”

Of the analysed species and subgroups, higher numbers of H. axyridis were recorded in Bt maize compared with non-Bt maize.

“Of those, only 2 showed significant effect sizes: higher numbers of H. axyridis and lower numbers of M. cingulum were recorded in Bt maize compared with non-Bt maize”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, positive effect sizes were observed for Anthocoridae.

“When untreated Bt maize was compared with pyrethroid treated non-Bt maize (mainly foliar and soil application), positive effect sizes (higher populations in Bt maize) were observed for Araneae, Cantharidae, Coccinellidae, Anthocoridae, and Cicadellidae and on higher taxonomic levels for Coleoptera, Hemiptera, and all taxa”

Positive effect sizes were present in Lepidoptera-active Bt maize for Anthocoridae recorded with visual counts.

“Positive effect sizes were present in Lepidoptera-active Bt maize for Coccinellidae collected with sticky traps and Anthocoridae recorded with visual counts and in stacked Bt maize for Neuroptera recorded with visual counts.”

Analyses with maize producing Coleoptera-active Bt proteins revealed no effects on any taxon.

“Analyses with maize producing Coleoptera-active Bt proteins revealed no effects on any taxon”

Lower populations in Bt maize were detected with beat cloths for Aphididae.

“Lower populations in Bt maize (negative effect sizes) were detected with sticky traps for Staphylinidae, Syrphidae, and Braconidae, and with beat cloths for Aphididae.”

Stacked maize producing Cry1A.105 & Cry2Ab & Cry3Bb revealed a negative effect size when all taxa were combined.

“and a negative effect size for maize producing Cry1A.105 & Cry2Ab & Cry3Bb when all taxa were combined”

Parasitism, evaluated by field-collected aphids, lacewing eggs, corn borer egg masses or corn borer larvae, was comparable between Bt maize (all Cry1Ab) and non-Bt maize (effect size −0.14 ± 0.15 [−0.45; 0.16], N = 19 records, 8 experiments, 6 articles).

“Parasitism, evaluated by field collected aphids, lacewing eggs, corn borer egg masses or corn borer larvae was comparable between Bt maize (all Cry1Ab) and non-Bt maize (effect size − 0.14 ± 0.15 [− 0.45; 0.16], N = 19 records, 8 experiments, 6 articles).”

Lower populations in Bt maize were detected with sticky traps for Syrphidae.

“Lower populations in Bt maize (negative effect sizes) were detected with sticky traps for Staphylinidae, Syrphidae, and Braconidae, and with beat cloths for Aphididae.”

Higher populations in Bt maize were observed with sticky traps for Coccinellidae.

“Higher populations in Bt maize (positive effect sizes) were observed with sticky traps for Coccinellidae, and with visual counts for Anthocoridae and Neuroptera.”

When only authors from public institutions and articles without declared private sector funding were selected, no difference between Bt maize and non-Bt maize was evident.

“When only authors from public institutions and articles without declared private sector funding were selected, no difference between Bt maize and non-Bt maize was evident”

Positive effect sizes were present in stacked Bt maize for Neuroptera recorded with visual counts.

“Positive effect sizes were present in Lepidoptera-active Bt maize for Coccinellidae collected with sticky traps and Anthocoridae recorded with visual counts and in stacked Bt maize for Neuroptera recorded with visual counts.”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, positive effect sizes were observed for all taxa combined.

“When untreated Bt maize was compared with pyrethroid treated non-Bt maize (mainly foliar and soil application), positive effect sizes (higher populations in Bt maize) were observed for Araneae, Cantharidae, Coccinellidae, Anthocoridae, and Cicadellidae and on higher taxonomic levels for Coleoptera, Hemiptera, and all taxa”

Inclusion of red-flagged records resulted in a similar outcome with a negative effect size for Braconidae.

“Inclusion of the red-flagged records resulted in a similar outcome of the analyses with a negative effect size (lower populations in Bt maize) for Staphylinidae, Syrphidae, Tachinidae, and Braconidae, and consequently for the higher taxonomic orders Diptera and Hymenoptera.”

No significant effect sizes were observed for all taxa combined with organophosphorous or microbial insecticides.

“No significant effect sizes were observed for Coleoptera and all taxa (organophosphorous or microbial).”

Predation rates, determined with sentinel corn borer egg masses or artificial prey, were similar between Bt and non-Bt maize (effect size −0.035 ± 0.14 [−0.31; 0.24], N = 14, 12 experiments, 5 articles).

“Predation rates, determined with sentinel corn borer egg masses or artificial prey, also were similar between Bt and non-Bt maize (effect size − 0.035 ± 0.14 [− 0.31; 0.24], N = 14, 12 experiments, 5 articles).”

Inclusion of red-flagged records resulted in a similar outcome with a negative effect size for Diptera.

“Inclusion of the red-flagged records resulted in a similar outcome of the analyses with a negative effect size (lower populations in Bt maize) for Staphylinidae, Syrphidae, Tachinidae, and Braconidae, and consequently for the higher taxonomic orders Diptera and Hymenoptera.”

Higher populations in Bt maize were observed with visual counts for Anthocoridae.

“Higher populations in Bt maize (positive effect sizes) were observed with sticky traps for Coccinellidae, and with visual counts for Anthocoridae and Neuroptera.”

Lower populations in Bt maize were detected with sticky traps for Braconidae.

“Lower populations in Bt maize (negative effect sizes) were detected with sticky traps for Staphylinidae, Syrphidae, and Braconidae, and with beat cloths for Aphididae.”

Higher populations in Bt maize were observed with visual counts for Neuroptera.

“Higher populations in Bt maize (positive effect sizes) were observed with sticky traps for Coccinellidae, and with visual counts for Anthocoridae and Neuroptera.”

For seed treatments with chloro-nicotinyl insecticides, positive effect sizes were observed for Araneae.

“Positive effect sizes were observed for Araneae, Hemiptera and the functional group of predators; no heterogeneity was detected for any analysed taxon or functional group.”

Data heterogeneity was present for Braconidae in analyses restricted to Lepidoptera-active Bt proteins.

“Data heterogeneity was present for Diptera, Syrphidae, Hymenoptera, and Braconidae.”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, positive effect sizes were observed for Coleoptera.

“When untreated Bt maize was compared with pyrethroid treated non-Bt maize (mainly foliar and soil application), positive effect sizes (higher populations in Bt maize) were observed for Araneae, Cantharidae, Coccinellidae, Anthocoridae, and Cicadellidae and on higher taxonomic levels for Coleoptera, Hemiptera, and all taxa”

No significant effect sizes were observed for Coleoptera with organophosphorous or microbial insecticides.

“No significant effect sizes were observed for Coleoptera and all taxa (organophosphorous or microbial).”

Inclusion of red-flagged records resulted in a similar outcome with a negative effect size for Staphylinidae.

“Inclusion of the red-flagged records resulted in a similar outcome of the analyses with a negative effect size (lower populations in Bt maize) for Staphylinidae, Syrphidae, Tachinidae, and Braconidae, and consequently for the higher taxonomic orders Diptera and Hymenoptera.”

Inclusion of red-flagged records resulted in a similar outcome with a negative effect size for Hymenoptera.

“Inclusion of the red-flagged records resulted in a similar outcome of the analyses with a negative effect size (lower populations in Bt maize) for Staphylinidae, Syrphidae, Tachinidae, and Braconidae, and consequently for the higher taxonomic orders Diptera and Hymenoptera.”

Sufficient records were available to analyse a total of 31 individual species and 8 other taxonomic subgroups.

“Sufficient records were available to analyse a total of 31 individual species and 8 other taxonomic subgroups.”

Analyses on functional group level revealed no effect of Bt maize on decomposers.

“Analyses on functional group level revealed no effect of Bt maize on decomposers, omnivores, and predators”

Meta-regressions showed no significant effect of plot size for any functional group.

“The meta-regressions showed no significant effect of plot size for any functional group (confidence interval of slope included 0 in all cases).”

Fail safe numbers for comparisons with untreated Bt and non-Bt maize indicate that the observed significances of effect sizes are not robust (fail safe number < 5n + 10) except for comparisons involving Braconidae.

“Fail safe numbers for comparisons with untreated Bt and non-Bt maize indicate that the observed significances of effect sizes are not robust (fail safe number < 5n + 10 [31]) except for the comparisons involving Braconidae”

• cites [31] Rosenberg MS. The file-drawer problem revisited: a general weighted method for calculating fail-safe numbers in meta-analysis. Evolution. 2005;59:464–8.Not checked yet

  Not checked in the free run.

The private-sector-associated negative effect derived mainly from Lepidoptera-active maize, stacked Coleoptera- and Lepidoptera-active maize, and the Bt protein Cry1Ab.

“This effect derived mainly from Lepidoptera-active maize and stacked Coleoptera- and Lepidoptera-active maize (Fig. 12, Additional file 7: Table S7.7), and the Bt protein Cry1Ab (Additional file 7: Table S7.8).”

Moderator analyses showed a positive effect of Cry1Ac maize on all taxa combined.

“Furthermore, the moderator analyses showed a positive effect of Cry1Ab maize on Anthocoridae and of Cry1Ac maize on all taxa combined”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, positive effect sizes were observed for Coccinellidae.

“When untreated Bt maize was compared with pyrethroid treated non-Bt maize (mainly foliar and soil application), positive effect sizes (higher populations in Bt maize) were observed for Araneae, Cantharidae, Coccinellidae, Anthocoridae, and Cicadellidae and on higher taxonomic levels for Coleoptera, Hemiptera, and all taxa”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, positive effect sizes were observed for Cicadellidae.

“When untreated Bt maize was compared with pyrethroid treated non-Bt maize (mainly foliar and soil application), positive effect sizes (higher populations in Bt maize) were observed for Araneae, Cantharidae, Coccinellidae, Anthocoridae, and Cicadellidae and on higher taxonomic levels for Coleoptera, Hemiptera, and all taxa”

Moderator analyses showed a positive effect of Cry1Ab maize on Anthocoridae.

“Furthermore, the moderator analyses showed a positive effect of Cry1Ab maize on Anthocoridae and of Cry1Ac maize on all taxa combined”

For seed treatments with chloro-nicotinyl insecticides, positive effect sizes were observed for the functional group of predators.

“Positive effect sizes were observed for Araneae, Hemiptera and the functional group of predators; no heterogeneity was detected for any analysed taxon or functional group.”

Analyses on functional group level revealed no effect of Bt maize on omnivores.

“Analyses on functional group level revealed no effect of Bt maize on decomposers, omnivores, and predators”

When untreated Bt maize was compared with pyrethroid-treated non-Bt maize, negative effect sizes were observed for Aphididae.

“In contrast, negative effect sizes (lower populations in Bt maize) were observed for Nitidulidae, Aphididae, and Formicidae.”

Analyses with only green-flagged records confirmed negative effect sizes for Hymenoptera.

“The negative effect sizes for Diptera, Syrphidae, and Hymenoptera were confirmed.”

The effect on Staphylinidae was not robust according to fail safe numbers and leave-one-out analyses.

“The effect was not robust according to fail safe numbers and “leave one out” analyses.”

When records for non-targets associated with Bt maize targeting Lepidoptera, Coleoptera, or stacks were examined together, a negative effect size on Staphylinidae was present, i.e., fewer rove beetles were recorded in Bt maize compared with non-Bt maize.

“When records for non-targets associated with Bt maize targeting Lepidoptera, Coleoptera, or stacks were examined together, a negative effect size on Staphylinidae was present, i.e., fewer rove beetles were recorded in Bt maize compared with non-Bt maize (Figs. 8, 10).”

The effects reported from meta-analyses derived from data of two articles with M. cingulum: Bruck et al. [37] (2 records) and Pilcher [34] (21 records).

“The effects reported from meta-analyses derived from data of two articles with M. cingulum: Bruck et al. [37] (2 records) and Pilcher [34] (21 records).”

• cites Bruck et al. [37] Bruck DJ, Lopez MD, Lewis LC, Prasifka JR, Gunnarson RD. Effects of transgenic Bacillus thuringiensis corn and permethrin on nontarget arthropods. J Agr Urban Entomol. 2006;23:111–24.Not checked yet

  Not checked in the free run.

• cites Pilcher [34] Pilcher CD. Phenological, physiological, and ecological influences of transgenic Bt corn on European corn borer management. Ph.D. thesis. Iowa State University, Ames, USA. 1999.Not checked yet

  Not checked in the free run.

Three articles [34-36] had a high influence on the significance of effect sizes.

“Three articles had a high influence on the significance of effect sizes [34–36].”

• cites [34] Pilcher CD. Phenological, physiological, and ecological influences of transgenic Bt corn on European corn borer management. Ph.D. thesis. Iowa State University, Ames, USA. 1999.Not checked yet

  Not checked in the free run.

• cites [35] Madrid JLC, Carrillo JLM, Martinez MBO, Pompa HAD, Escobedo JA, Quinones FJ, et al. Transportability of non-target arthropod field data for the use in environmental risk assessment of genetically modified maize in Northern Mexico. J Appl Entomol. 2018;142:525–38. sourceNot checked yet

  Not checked in the free run.

• cites [36] Szenasi A, Palinkas Z, Zalai M, Schmitz OJ, Balog A. Short-term effects of different genetically modified maize varieties on arthropod food web properties: an experimental field assessment. Sci Rep. 2014;4:5315. sourceNot checked yet

  Not checked in the free run.

The positive effect size for Araneae in Bt maize compared with chloro-nicotinyl-treated non-Bt maize remained significant in all leave-one-out analyses.

“The positive effect size for Araneae in Bt maize compared with chloro-nicotinyl-treated non-Bt maize remained significant in all “leave one out” analyses.”

The study by Szenasi et al. [36] was mainly responsible for the negative effect size seen with sticky traps.

“The study by Szenasi et al. [36] was mainly responsible for the negative effect size seen with sticky traps.”

• cites Szenasi et al. [36] Szenasi A, Palinkas Z, Zalai M, Schmitz OJ, Balog A. Short-term effects of different genetically modified maize varieties on arthropod food web properties: an experimental field assessment. Sci Rep. 2014;4:5315. sourceNot checked yet

  Not checked in the free run.

More significant effect sizes on different taxa (8 of 20) were evident when untreated Bt maize was compared with alternative methods (pyrethroids and other insecticides) employed for pest control in non-Bt maize.

“More significant effect sizes on different taxa (8 of 20) were evident when untreated Bt maize was compared with alternative methods (pyrethroids and other insecticides) employed for pest control in non-Bt maize”

Main meta-analyses revealed no significant effect sizes for 36 of 39 analyzed taxa when both Bt maize and non-Bt maize were not treated with insecticides.

“Our main meta-analyses revealed no significant effect sizes for 36 of 39 analyzed taxa (Fig. 8, Additional file 7: Table S7.1) when both Bt maize and non-Bt maize were not treated with insecticides.”

The meta-analyses showed a reduced number of Nitidulidae in Bt maize when compared with pyrethroid-treated non-Bt maize.

“Our meta-analyses also showed a reduced number of Nitidulidae in Bt maize when compared with pyrethroid-treated non-Bt maize (Fig. 14).”

When non-Bt maize was treated with pyrethroids, fail safe numbers indicate robustness of positive effect sizes for all taxa, Araneae, Hemiptera, and predators.

“When non-Bt maize was treated with pyrethroids, fail safe numbers indicate robustness of positive effect sizes (higher populations in untreated Bt maize compared to insecticide-treated non-Bt maize) for all taxa, Araneae, Hemiptera, and predators.”

There were no effects on the staphylinid species Aleochara bipustulata and Anotylus rugosus when analyzed individually.

“There also were no effects on the staphylinid species Aleochara bipustulata and Anotylus rugosus when analyzed individually (Fig. 13).”

When non-Bt maize was treated with chloro-nicotinyl, microbial insecticides, or organophosphorous, fail safe numbers indicate that significant effects were not robust.

“When non-Bt maize was treated with chloro-nicotinyl, microbial insecticides, or organophosphorous, fail safe numbers indicate that significant effects were not robust.”

The effect on Tachinidae was not robust according to fail safe numbers.

“The effect, however, was not robust according to fail safe numbers.”

Tachinidae were less abundant in Lepidoptera-active Bt maize.

“The second group of parasitoids of O. nubilalis is Tachinidae, which also were less abundant in Lepidoptera-active Bt maize.”

The effect on Nitidulidae was not robust according to fail safe numbers and leave-one-out analyses.

“The effect, however, was not robust according to fail safe numbers and “leave one out” analyses.”

Studies using Bt products showed no significant effect size.

“while studies using Bt products showed no significant effect size.”

Non-robust negative effect sizes were seen in studies with private sector contribution.

“Non-robust negative effect sizes were seen in studies with private sector contribution (authors and/or funding from private sector product developers)”

Higher populations of Anthocoridae were recorded in Lepidoptera-active Bt maize (Cry1Ab, visual counts).

“In contrast, higher populations of Anthocoridae (flower bugs) were recorded in Lepidoptera-active Bt maize (Cry1Ab, visual counts) and higher populations of Neuroptera (lacewings) were recorded in Bt maize with visual counts (of all stages) and when only eggs were analyzed.”

The analysis with stacked maize was dominated by two articles, [35] (68 of 120 records) and [36] (28 records).

“The analysis with stacked maize was dominated by two articles, [35] (68 of 120 records) and [36] (28 records).”

• cites [35] Madrid JLC, Carrillo JLM, Martinez MBO, Pompa HAD, Escobedo JA, Quinones FJ, et al. Transportability of non-target arthropod field data for the use in environmental risk assessment of genetically modified maize in Northern Mexico. J Appl Entomol. 2018;142:525–38. sourceNot checked yet

  Not checked in the free run.

• cites [36] Szenasi A, Palinkas Z, Zalai M, Schmitz OJ, Balog A. Short-term effects of different genetically modified maize varieties on arthropod food web properties: an experimental field assessment. Sci Rep. 2014;4:5315. sourceNot checked yet

  Not checked in the free run.

Negative and non-robust effect sizes were obtained for Nitidulidae, Aphididae, and Formicidae.

“In contrast, negative and non-robust effect sizes (fewer individuals recorded in Bt maize) were obtained for Nitidulidae, Aphididae, and Formicidae.”

No Bt effects were evident in studies without private sector contribution.

“while no Bt effects were evident in studies without such private sector contribution (Fig. 12).”

Predatory species were harmed by insecticide treatments.

“In particular predatory species were harmed by insecticide treatments.”

When red-flagged records were excluded, the upper boundary of the confidence interval for Syrphidae was 0.0000.

“When red flagged records were excluded, the upper boundary of the confidence interval for Syrphidae was 0.0000 (Fig. 8).”

One study [34] dominated the analysis of all-green-flagged records as it contributed 20 of 38 records.

“One study [34] dominated the analysis of all-green flagged records as it contributed 20 of 38 records.”

• cites [34] Pilcher CD. Phenological, physiological, and ecological influences of transgenic Bt corn on European corn borer management. Ph.D. thesis. Iowa State University, Ames, USA. 1999.Not checked yet

  Not checked in the free run.

Two articles [34] and [35] were mainly responsible for the negative effect size in the private sector contribution group, contributing more than half of the records.

“Two articles were mainly responsible for the negative effect size as they contributed more than half of the records in the group of private sector contribution: [34] and [35].”

• cites [34] Pilcher CD. Phenological, physiological, and ecological influences of transgenic Bt corn on European corn borer management. Ph.D. thesis. Iowa State University, Ames, USA. 1999.Not checked yet

  Not checked in the free run.

• cites [35] Madrid JLC, Carrillo JLM, Martinez MBO, Pompa HAD, Escobedo JA, Quinones FJ, et al. Transportability of non-target arthropod field data for the use in environmental risk assessment of genetically modified maize in Northern Mexico. J Appl Entomol. 2018;142:525–38. sourceNot checked yet

  Not checked in the free run.

Untreated Bt maize harboured generally more arthropods than pyrethroid-treated non-Bt maize (all taxa combined).

“Untreated Bt maize harboured generally more arthropods than pyrethroid treated non-Bt maize (all taxa combined, Fig. 14).”

Positive, non-robust effect sizes were present for beetles (Coleoptera), in particular Cantharidae and Coccinellidae, Anthocoridae and Cicadellidae.

“Positive, non-robust effect sizes were present for beetles (Coleoptera), in particular Cantharidae and Coccinellidae, Anthocoridae and Cicadellidae.”

Most records for insecticide-treated non-Bt maize were for pyrethroids mainly applied as foliar sprays, sometimes also as soil insecticides, and for seed treatments with chloro-nicotinyl.

“Most of the records for insecticide-treated non-Bt maize were for pyrethroids that were mainly applied as foliar sprays, sometimes also as soil insecticides, and for seed treatments with chloro-nicotinyl.”

No significance was obtained for Syrphidae when Bt maize targeting different orders or Bt proteins were analysed separately.

“However, no significance was obtained when Bt maize targeting different orders or Bt proteins were analysed separately (Fig. 9).”

On the species level, a weak positive effect was observed for H. axyridis.

“On the species level, a weak positive effect was observed for H. axyridis (Fig. 13).”

One study [34] with Cry1Ab-producing maize (21 of 63 records, red flags excluded, all target orders) had a high influence on the Syrphidae analyses with mostly negative effect sizes.

“One study [34] with Cry1Ab-producing maize (21 of 63 records, red flags excluded, all target orders) had a high influence on those analyses with mostly negative effect sizes.”

• cites [34] Pilcher CD. Phenological, physiological, and ecological influences of transgenic Bt corn on European corn borer management. Ph.D. thesis. Iowa State University, Ames, USA. 1999.Not checked yet

  Not checked in the free run.

Seed treatments with chloro-nicotinyl in non-Bt maize resulted in higher populations in untreated Bt maize for spiders (Araneae), Hemiptera, and predators, with positive but non-robust effect sizes.

“Seed treatments with chloro-nicotinyl in non-Bt maize also resulted in higher populations in the untreated Bt maize (positive, non-robust effect sizes) for spiders (Araneae), Hemiptera, and for the functional group of predators (Fig. 14).”

For microbial insecticides, the effect can be linked to one study that used spinosad.

“In the case of microbial insecticides, the effect can be linked to one study that used spinosad [58]”

• cites [58] Musser FR, Shelton AM. Bt sweet corn and selective insecticides: Impacts on pests and predators. J Econ Entomol. 2003;96:71–80. sourceNot checked yet

  Not checked in the free run.

Most records had high or medium internal and external validity (green or yellow flag).

“Most records had high or medium internal and external validity (green or yellow flag).”

Only few studies had a green flag for all appraisal criteria.

“Nevertheless, only few studies had a green flag for all appraisal criteria (Fig. 10B).”

Information on field management was often incomplete.

“We also observed that information on field management was often incomplete”

Our field data showed no effects for 36 analyzed taxonomic groups.

“Our field data, showing no effects for 36 analyzed taxonomic groups”

Verification of Bt protein expression in the plant was often not considered by the study-authors.

“In particular, verification of Bt protein expression in the plant was often not considered by the study-authors.”

A number of relevant studies could only be summarized narratively, but not used for quantitative analyses.

“Consequently, a number of relevant studies could only be summarized narratively (Additional file 3: Table S3.5), but not used for quantitative analyses.”

Most non-target studies were conducted in replicated plots within a field.

“Most non-target studies were conducted in replicated plots within a field.”

Meta-regressions suggested a general lack of effect of plot size on outcomes at the level of functional guild.

“Although our meta-regressions suggested a general lack of effect of plot size on outcomes at the level of functional guild”

Some articles contributed high numbers of records as they reported on experiments with multiple transformation events from several locations and years.

“In fact, some articles (e.g., [34–36]) contributed high numbers of records as they reported on experiments with multiple transformation events from several locations and years.”

• cites [34] Pilcher CD. Phenological, physiological, and ecological influences of transgenic Bt corn on European corn borer management. Ph.D. thesis. Iowa State University, Ames, USA. 1999.Not checked yet

  Not checked in the free run.

• cites [35] Madrid JLC, Carrillo JLM, Martinez MBO, Pompa HAD, Escobedo JA, Quinones FJ, et al. Transportability of non-target arthropod field data for the use in environmental risk assessment of genetically modified maize in Northern Mexico. J Appl Entomol. 2018;142:525–38. sourceNot checked yet

  Not checked in the free run.

• cites [36] Szenasi A, Palinkas Z, Zalai M, Schmitz OJ, Balog A. Short-term effects of different genetically modified maize varieties on arthropod food web properties: an experimental field assessment. Sci Rep. 2014;4:5315. sourceNot checked yet

  Not checked in the free run.

Some groups of non-target species, such as bees and other pollinators, butterflies and moths, aquatic organisms, or vertebrates are usually not studied in such experimental setups.

“Some groups of non-target species, such as bees and other pollinators, butterflies and moths, aquatic organisms, or vertebrates are usually not studied in such experimental setups.”

The sample size in Bt and non-Bt plots or fields was in most cases similar.

“Bt and non-Bt plots were sampled with the same methodology (ES), and the sample size in Bt and non-Bt plots or fields was in most cases similar (SS).”

Authors generally reported that they used near-isolines or nearest comparators for Bt and non-Bt maize.

“Authors generally reported that they used near-isolines or nearest comparators for Bt and non-Bt maize (RC).”

The time of collections was suitable to cover the population over the season.

“Generally, invertebrates were collected with common and appropriate methods (MS) and also the time of collections was suitable to cover the population over the season (MT).”

Usually, three or more plots per treatment or four or more separated fields were deployed.

“Usually, three or more plots per treatment or four or more separated fields were deployed (RE).”

Bt and non-Bt plots were sampled with the same methodology.

“Bt and non-Bt plots were sampled with the same methodology (ES), and the sample size in Bt and non-Bt plots or fields was in most cases similar (SS).”

Invertebrates were collected with common and appropriate methods.

“Generally, invertebrates were collected with common and appropriate methods (MS) and also the time of collections was suitable to cover the population over the season (MT).”

In more than 70% of the records, at least four samples were performed over the season.

“In more than 70% of the records, ≥ four samples were performed over the season (SD).”

Reusing control data without correction for multiple tests may make the results slightly less conservative.

“This situation is analogous to the use of Dunnett’s test, where each treatment is compared with the same control, but without correction for multiple tests. Thus, our results here might be slightly less conservative.”

The design of the studies consisted in most cases of plots, which were randomized or systematically distributed across the field without evidence of clustering.

“The design of the studies consisted in most cases of plots, which were randomized or systematically distributed across the field without evidence of clustering (FC, 2-letter codes used in Fig. 7 and in Ref. [25]).”

There was no indication that Bt and non-Bt fields had a different history of management before the study.

“There was no indication that Bt and non-Bt fields had a different history of management before the study (HM).”

There also was no indication that Bt and non-Bt plots received different pesticides during the study, except for insecticide treatments that were part of the experimental design.

“There also was no indication that Bt and non-Bt plots received different pesticides (e.g., fungicides, herbicides) during the study, except for insecticide treatments that were part of the experimental design (PD).”

Laboratory studies do not support the hypothesis of direct effects of Bt maize on Staphylinidae.

“Laboratory studies do not support the hypothesis of direct effects of Bt maize on Staphylinidae.”

The overall effect on Staphylinidae cannot be linked to particular types of Bt maize, Bt proteins, or species.

“In contrast to the results for parasitoids of target Lepidoptera pests and for Nitidulidae, the overall effect on Staphylinidae thus cannot be linked to particular types of Bt maize, Bt proteins, or species.”

Siegfried et al. [38] observed reduced parasitism rates of diapausing O. nubilalis larvae collected in commercial Bt maize (event 176) compared with non-Bt maize fields by M. cingulum or Eriborus terebrans.

“In addition, Siegfried et al. [38] observed reduced parasitism rates of diapausing O. nubilalis larvae collected in commercial Bt maize (event 176) compared with non-Bt maize fields by M. cingulum or Eriborus terebrans (Hym.: Ichneumonidae).”

• cites Siegfried et al. [38] Siegfried BD, Zoerb AC, Spencer T. Development of European corn borer larvae on event 176 Bt corn: influence on survival and fitness. Entomol Exp Appl. 2001;100:15–20. sourceNot checked yet

  Not checked in the free run.

Records from Madrid et al. [35] may have contributed to the overall negative effect on Tachinidae.

“Records from Madrid et al. [35], who collected Tachinidae in Cry1A.105 & Cry2Ab producing maize in Brazil, also may have contributed to the overall negative effect on this family.”

• cites Madrid et al. [35] Madrid JLC, Carrillo JLM, Martinez MBO, Pompa HAD, Escobedo JA, Quinones FJ, et al. Transportability of non-target arthropod field data for the use in environmental risk assessment of genetically modified maize in Northern Mexico. J Appl Entomol. 2018;142:525–38. sourceNot checked yet

  Not checked in the free run.

No clear relationship between Bt status and the presence of sap beetles in pheromone-baited traps was reported by Dowd [52].

“However, no clear relationship between Bt status and the presence of sap beetles in pheromone baited traps was reported by Dowd [52].”

• cites Dowd [52] Dowd PF. Dusky sap beetles (Coleoptera: Nitidulidae) and other kernel damaging insects in Bt and non-Bt sweet corn in Illinois. J Econ Entomol. 2000;93:1714–20. sourceNot checked yet

  Not checked in the free run.

Whenever measures are taken to reduce target pests, e.g., by biological control, pesticides, or Bt crops, an impact on the natural enemy community of those pests is expected [46].

“Whenever measures are taken to reduce target pests, e.g., by biological control, pesticides, or Bt crops, an impact on the natural enemy community of those pests is expected [46].”

• cites [46] Shelton AM, Zhao JZ, Roush RT. Economic, ecological, food safety, and social consequences of the deployment of Bt transgenic plants. Ann Rev Entomol. 2002;47:845–81. sourceNot checked yet

  Not checked in the free run.

Bourguet et al. [42] collected O. nubilalis larvae and recorded reduced parasitism by Lydella thompsoni and Pseudoperichaeta nigrolineata in Cry1Ab-producing Bt maize (event 176) in France.

“Bourguet et al. [42] collected O. nubilalis larvae and recorded reduced parasitism by Lydella thompsoni and Pseudoperichaeta nigrolineata (both Diptera: Tachinidae) in Cry1Ab-producing Bt maize (event 176) in France.”

• cites Bourguet et al. [42] Bourguet D, Chaufaux J, Micoud A, Delos M, Naibo B, Bombarde F, et al. Ostrinia nubilalis parasitism and the field abundance of non-target insects in transgenic Bacillus thuringiensis corn (Zea mays). Environ Biosaf Res. 2002;1:49–60. sourceNot checked yet

  Not checked in the free run.

The number of M. cingulum adults captured in non-Bt plots in another study was double the number captured in Bt plots [41].

“Similarly, the number of M. cingulum adults captured in non-Bt plots in another study was double the number captured in Bt plots [41]”

• cites [41] Sked SL, Calvin DD. Synchronization of Macrocentrus cingulum (Reinhardt) (Hymenoptera: Braconidae) with its host and the effects of transgenic crop systems on this natural enemy. In: Hoddle MS, editor. California conference on biological control III; 15–16 August, 2002. Berkeley: Center for Biological Control, College of Natural Resources, University of California; 2002. p. 16–22.Not checked yet

  Not checked in the free run.

More non-target invertebrates were recorded in untreated Bt maize compared to pyrethroid-treated non-Bt maize.

“More non-target invertebrates, however, were recorded in untreated Bt maize compared to pyrethroid-treated non-Bt maize.”

• cites [10] Marvier M, McCreedy C, Regetz J, Kareiva P. A meta-analysis of effects of Bt cotton and maize on nontarget invertebrates. Science. 2007;316:1475–7. sourceNot checked yet

  Not checked in the free run.

The observed pattern of non-robust negative effect sizes in private-sector-contributed studies does not support the hypothesis that the private sector hides adverse effects while public-sector scientists are more interested in publishing effects.

“This does not support the hypotheses that the private sector might have vested interests in hiding (adverse) effects data while public sector scientists might be more interested in publishing effects.”

Previous meta-analyses came to the same conclusion about insecticide-treated non-Bt maize comparisons.

“Previous meta-analyses came to the same conclusion [13, 14].”

• cites [13] Wolfenbarger LL, Naranjo SE, Lundgren JG, Bitzer RJ, Watrud LS. Bt crop effects on functional guilds of non-target arthropods: a meta-analysis. PLoS ONE. 2008;3: e2118. sourceNot checked yet

  Not checked in the free run.

• cites [14] Naranjo SE. Impact of Bt crops on non-target invertebrates and insecticide use patterns. CAB Rev. 2009;4:11.Not checked yet

  Not checked in the free run.

It is likely that the weak negative effect size observed for staphylinids in the meta-analyses may have derived from indirect, food web related effects, other ecological factors, or from chance.

“It is thus likely that the weak negative effect size observed for staphylinids in our meta-analyses may have derived from indirect, food web related effects, other ecological factors, or from chance.”

Direct toxicity of the plant-produced Bt proteins is unlikely.

“Direct toxicity of the plant-produced Bt proteins is thus unlikely.”

Negative effects of Bt maize on Braconidae, Tachinidae, and Nitidulidae are likely linked to reduced numbers of corn borer larvae in Lepidoptera-active maize.

“Negative effects of Bt maize observed in some meta-analyses on Braconidae, Tachinidae, and Nitidulidae are likely linked to reduced numbers of corn borer larvae in Lepidoptera-active maize”

Bt maize compared to untreated non-Bt maize reduced parasitoid abundance.

“Follow-up analyses of the same dataset by functional guilds confirmed that Bt maize (data for Cry1Ab- and Cry3Bb-producing maize combined) compared to untreated non-Bt maize reduced parasitoid abundance”

• cites [13] Wolfenbarger LL, Naranjo SE, Lundgren JG, Bitzer RJ, Watrud LS. Bt crop effects on functional guilds of non-target arthropods: a meta-analysis. PLoS ONE. 2008;3: e2118. sourceNot checked yet

  Not checked in the free run.

Overall effects of Cry3Bb-producing maize were insignificant, either without or with pyrethroid application in non-Bt maize.

“Overall effects of Cry3Bb-producing maize were insignificant, either without or with pyrethroid application in non-Bt maize [10].”

• cites [10] Marvier M, McCreedy C, Regetz J, Kareiva P. A meta-analysis of effects of Bt cotton and maize on nontarget invertebrates. Science. 2007;316:1475–7. sourceNot checked yet

  Not checked in the free run.

Our field data largely confirm the laboratory data.

“Our field data, showing no effects for 36 analyzed taxonomic groups, thus largely confirm the laboratory data.”

Previous meta-analyses selected data on the finest available taxonomic level.

“Previous meta-analyses [10, 12–14] selected data on the finest available taxonomic level.”

• cites [10] Marvier M, McCreedy C, Regetz J, Kareiva P. A meta-analysis of effects of Bt cotton and maize on nontarget invertebrates. Science. 2007;316:1475–7. sourceNot checked yet

  Not checked in the free run.

• cites [12] Duan JJ, Lundgren JG, Naranjo S, Marvier M. Extrapolating non-target risk of Bt crops from laboratory to field. Biol Lett. 2010;6:74–7. sourceNot checked yet

  Not checked in the free run.

• cites [13] Wolfenbarger LL, Naranjo SE, Lundgren JG, Bitzer RJ, Watrud LS. Bt crop effects on functional guilds of non-target arthropods: a meta-analysis. PLoS ONE. 2008;3: e2118. sourceNot checked yet

  Not checked in the free run.

• cites [14] Naranjo SE. Impact of Bt crops on non-target invertebrates and insecticide use patterns. CAB Rev. 2009;4:11.Not checked yet

  Not checked in the free run.

Despite efforts to achieve data independence, data from control plots were used multiple times if different Bt maize lines had the same control, so a certain level of data-dependence remained.

“Despite this effort to achieve data independence, data from control plots were used multiple times if different Bt maize lines had the same control, so a certain level of data-dependence remained.”

In Comas et al.'s meta-analysis of 13 Bt maize field trials in Spain, none of the analysed taxa for each sampling method showed effect sizes significantly different from zero.

“None of the analysed taxa for each sampling method (7 taxa for visual sampling, 7 for pitfalls, 12 for sticky traps) showed effect sizes significantly different from zero.”

• cites [15] Comas C, Lumbierres B, Pons X, Albajes R. No effects of Bacillus thuringiensis maize on nontarget organisms in the field in southern Europe: a meta-analysis of 26 arthropod taxa. Transgenic Res. 2014;23:135–43. sourceNot checked yet

  Not checked in the free run.

Negative effect sizes for Staphylinidae and Syrphidae and positive effect sizes for Anthocoridae, Coccinellidae, and Neuroptera remain unexplained.

“while negative effect sizes for Staphylinidae and Syrphidae and positive effect sizes for Anthocoridae, Coccinellidae, and Neuroptera remain unexplained.”

Krogh et al.'s systematic review found no effects on soil invertebrates across all crops and Bt proteins.

“Over all crops and Bt proteins, no effects on soil invertebrates were detected.”

• cites [17] Krogh PH, Kostov K, Damgaard CF. The effect of Bt crops on soil invertebrates: a systematic review and quantitative meta-analysis. Transgenic Res. 2020;29:487–98. sourceNot checked yet

  Not checked in the free run.

Except for Staphylinidae, no controlled exposure studies examining direct toxic effects of Bt proteins are available for taxa where our meta-analyses showed lower populations in Bt maize.

“With the exception of Staphylinidae [53–55], no controlled exposure studies to examine direct toxic effects of Bt proteins are available for the taxa where our meta-analyses showed lower populations in Bt maize.”

• cites [53] Garcia M, Ortego F, Castañera P, Farinós G. Effects of exposure to the toxin Cry1Ab through Bt maize fed-prey on the performance and digestive physiology of the predatory rove beetle Atheta coriaria. Biol Control. 2010;55:225–33. sourceNot checked yet

  Not checked in the free run.

• cites [54] García M, Ortego F, Castañera P, Farinós G. Assessment of prey-mediated effects of the coleopteran-specific toxin Cry3Bb1 on the generalist predator Atheta coriaria (Coleoptera: Staphylinidae). Bull Ent Res. 2012;102:293–302. sourceNot checked yet

  Not checked in the free run.

• cites [55] Porcar M, García-Robles I, Domínguez-Escribà L, Latorre A. Effects of Bt Cry1Ab and Cry3Aa endotoxins on predatory Coleoptera tested thourgh artificial diet incorporation assays. Bull Ent Res. 2010;100:297–302. sourceNot checked yet

  Not checked in the free run.

Reported adverse effects of Bt proteins or Bt plant material on non-target species in laboratory studies appeared to be linked to natural enemies being provided with sublethally affected hosts or prey.

“Putative adverse effects of Bt proteins or Bt plant material on non-target species have been reported in a review of laboratory studies [64], but such effects appeared to be linked to natural enemies being provided with sublethally affected hosts or prey [65].”

• cites [64] Lövei GL, Andow DA, Arpaia S. Transgenic insecticidal crops and natural enemies: a detailed review of laboratory studies. Environ Entomol. 2009;38:293–306. sourceNot checked yet

  Not checked in the free run.

• cites [65] Shelton AM, Naranjo SE, Romeis J, Hellmich RL, Wolt J, Federici BA, Albajes R, Bigler F, Burgess EPJ, Dively GP, Gatehouse AMR, Malone LA, Roush R, Sears M, Sehnal F. Setting the record straight: a rebuttal to an erroneous analysis on transgenic insecticidal crops and natural enemies. Transgenic Res. 2009;18:317–22. sourceNot checked yet

  Not checked in the free run.

Effects on predators, herbivores, omnivores, detritivores and mixed feeders were insignificant in the follow-up analyses.

“while effects on predators, herbivores, omnivores, detritivores and mixed feeders were insignificant [13].”

• cites [13] Wolfenbarger LL, Naranjo SE, Lundgren JG, Bitzer RJ, Watrud LS. Bt crop effects on functional guilds of non-target arthropods: a meta-analysis. PLoS ONE. 2008;3: e2118. sourceNot checked yet

  Not checked in the free run.

The conclusions of this review are largely consistent with conclusions drawn in previous meta-analyses.

“The conclusions of our work are largely consistent with the conclusions drawn previously.”

Data availability in general was a major drawback for the construction of the database.

“Data availability in general was a major drawback for the construction of the database.”

Depending on the mobility of the studied taxa, extrapolating conclusions from small-scale plot-experiments to actual farmer fields might be limited.

“Depending on the mobility of the studied taxa, extrapolating conclusions from such small-scale plot-experiments to actual farmer fields might be limited [68, 69].”

• cites [68] Andow DA, Zwahlen C. Assessing environmental risks of transgenic plants. Ecol Lett. 2005;9:196–214. sourceNot checked yet

  Not checked in the free run.

• cites [69] Jepson PC, Sherratt TN. The dimensions of space and time in the assessment of ecotoxicological risks. In: Baird DJ, Maltby L, Greig-Smith PW, Douben PET, editors. Ecotoxicology: ecological dimensions, vol. 3. London: Chapman and Hall; 1996. p. 43–54. sourceNot checked yet

  Not checked in the free run.

This systematic review generally supports the conclusions of regulatory environmental risk assessments that no unacceptable risks for non-target organisms and biodiversity exist.

“Our systematic review generally supports these conclusions.”

One shortcoming of this review and all prior ones is the limited availability of appropriate data, with the consequence that some data had to be estimated for quantitative analyses, while other datasets could not be analyzed at all.

“One shortcoming of this current review and all prior ones, however, is the limited availability of appropriate data (in particular raw data) with the consequence that some data had to be estimated (rather than calculated) for the quantitative analyses, while other datasets could not be analyzed at all.”

The conclusions of this review are largely limited to invertebrates inhabiting maize fields.

“The conclusions of our review are thus largely limited to invertebrates inhabiting maize fields.”

Many individual statistical tests were conducted with different datasets without correcting confidence limits for multiple testing.

“It has to be noted that we conducted many individual statistical tests with different datasets without correcting confidence limits for multiple testing.”

South America represents approximately one third of global Bt maize production.

“South America is particularly underrepresented as it represents approximately one third of the global Bt maize production [66].”

• cites [66] Naranjo SE, Hellmich RL, Romeis J, Shelton AM, Velez AM. The role and use of genetically engineered insect-resistant crops in IPM systems. In: Kogan M, Heinrichs E, editors. Integrated management of insect pests: current and future developments. Cambridge: Burleigh Dodds Science Publishing; 2020. p. 283–340.Not checked yet

  Not checked in the free run.

All commercially cultivated Bt maize transformation events have gone through regulatory environmental risk assessments that concluded no unacceptable risks for non-target organisms and biodiversity.

“The Bt maize transformation events that have been cultivated commercially worldwide have all gone through regulatory environmental risk assessments that concluded that no unacceptable risks for non-target organisms and biodiversity exist [2, 73, 74].”

• cites [2] ISAAA. GM approval database. International service for the acquisition of agri-biotech applications (ISAAA). 2019. www.isaaa.org/gmapprovaldatabase. Accessed 26 Apr 2022.Not checked yet

  Not checked in the free run.

• cites [73] Craig W, Tepfer M, Degrassi G, Ripandelli D. An overview of general features of risk assessments of genetically modified crops. Euphytica. 2008;164:853–80. sourceNot checked yet

  Not checked in the free run.

• cites [74] Secretariat of the Convention on Biological Diversity. Cartagena protocol on biosafety to the convention on biological diversity: text and annexes. Montreal, Canada. 2000. https://bch.cbd.int/protocol/cpb_publications.shtml. Accessed 22 Apr 2022.Not checked yet

  Not checked in the free run.