Dam vaccination for the prevention of neonatal diarrhea caused by enterotoxigenic E. coli in calves-a systematic review and meta-analysis

Highlights

• •Vaccination of dams for the prevention of Escherichia coli prevents death of neonatal calves.
• •Homologous and heterologous challenge trials showed strong evidence for efficacy.
• •Limited data is available from field trials, showing no advantage of vaccination.
• •Generalized conclusions from field trial data are difficult to make.
• •Biased results cannot be ruled out as few mention randomization or blinding.

Abstract

Diarrhea in neonatal calves caused by enterotoxigenic Escherichia coli (ETEC) is a disease that negatively affects the welfare and production of cattle. We present a review of all literature published in 1950 and later, evaluating vaccine effectiveness and efficacy in protecting calves from diarrhea and death from infection with ETEC by vaccinating dams during gestation. Of 3677 citations identified, 61 were eligible for full text review. Data was extracted from 22 citations and separated into 3 subgroups: field trials, heterologous challenge trials, and homologous challenge trials. Field trials showed no evidence to support that vaccinating dams decreased the rate of death or diarrhea in calves under 14 days of age, but due to a limited number and quality of field trials that met our inclusion criteria, it is difficult to draw general conclusions from this finding. However, in both homologous and heterologous challenge trials, vaccines helped prevent death in study calves. The outcome diarrhea was eliminated from analysis in challenge trials due to unexplained statistical heterogeneity which may be caused by a variety of definitions for diarrhea. Very few studies included information on randomization, blinding, or funding sources. Based on funnel plots, some indication for publication bias exists for heterologous challenge trials, however, a sensitivity analysis using the trim and fill method did not change results. There is evidence for the efficacy of ETEC vaccination of dams in the prevention of neonatal calf death.

1. Introduction

Diarrhea in calves, also known as scours, is a common disease that negatively affects the cattle industry. The National Animal Health Monitoring System reports that 32 % of calf mortality on dairies (United States Department of Agriculture, 2021) and 11.9 % on cow-calf operations in calves less than 3 weeks old (United States Department of Agriculture, 2020) is caused by digestive problems including diarrhea. Diarrhea can affect calves at various ages, between days to months old, often depending on the causative pathogen, and can cause substantial economic loss in cattle production apart from its welfare implications. Enterotoxigenic E. coli (ETEC) are bacteria that cause diarrhea in young calves less than a week old. Infection happens through fecal-oral transmission. Diarrhea caused by this pathogen can cause substantial morbidity and mortality due to severe dehydration and electrolyte imbalance in affected calves (Dubreuil, 2016).

Prevention of calf diarrhea can be managed through multiple practices, including immunization by vaccination of the pregnant dam. Fimbriae, also called pili, are filaments on the surface of bacteria. In ETEC, fimbriae are virulence factors that mediate adhesion and colonization of E. coli to intestinal cells during diarrhea events. In calves, F5 (formerly known as K99) and F41 are the most important type of fimbriae and also the main immunogenic proteins used in vaccines (Dubreuil et al., 2016). Serotypes of E. coli are defined by the structure of the O-antigen on the lipopolysaccharide layer of the bacterial outer membrane, e.g. O8, O9, O101 but are not directly responsible for adhesion and toxin production.

The objective of this study was to evaluate the efficacy and effectiveness of commercial and experimental vaccines in preventing diarrhea and death in calves caused by ETEC from research previously published on this topic through a systematic review and meta-analysis. To our knowledge, there are no previous such studies. Since results from individual studies are not always in agreement, summarizing the evidence in this manner may be helpful in evaluating if vaccination of dams is effective in preventing morbidity and mortality in newborn calves from diarrhea caused by ETEC, while also taking into consideration the quality of the available studies.

2. Materials and methods

2.1. Study protocol and registration

The study protocol was deposited in escholarship.org, the online repository of the University of California, and registered with Systematic Reviews for Animals and Food (SYREAF) (http://www.syreaf.org) in April 2023. We followed PRISMA-P guidelines for Systematic review and Meta-Analysis Protocols in preparing the protocol (Shamseer, 2015).

2.2. Literature search and other sources of information

Search results from a previous scoping review on vaccination for the prevention of neonatal calf diarrhea in cow-calf operations were used and complemented with an updated search using the same search criteria (Maier, 2022). We eliminated articles with target pathogens for vaccination other than ETEC. The database search was conducted through the Carlson Heath Library at the University of California, Davis. The original search for the scoping review was conducted in 2020 using databases Medline using the interface (Pubmed interface, 1972–2020), CAB abstracts (CAB Direct interface, 1972–2020), and Biosis (Web of Science interface, 1926–2020) in November 2019 and was updated with follow-up searches for publications in March 2023 and March 2025 using the same search algorithm. See supplementary file S1 for full search criteria. The results were transferred into Endnote (Endnote, Clarivate Analytics, Philadelphia, Pennsylvania) and duplicates were deleted. Conference proceedings and vaccine manufacturers’ unpublished literature were searched and evaluated as part of the original search. Finally, we also searched the USDA’s Center for Veterinary Biologics website for appropriate studies.

2.3. Eligibility criteria

In order for studies to be included they had to meet the following requirements: published in 1950 or later as defined in the original scoping review (Maier et al., 2022), published in English, original scientific reports, all study designs except case studies or case series, more than 500 words and peer reviewed, no restrictions based on cattle use/commodity. We decided to also include results published through the Center for Veterinary Biologics at the United States Department of Food and Agriculture (USDA) as they represent valuable information on commercial vaccine testing. Studies had to include dams vaccinated against ETEC and have a diagnosis of ETEC through laboratory diagnosis or clinical diagnosis after receiving E. coli challenge exposure to ensure a case definition of colibacillosis. Our two outcomes of interest were diarrhea caused by ETEC or death in study calves. As one of the inclusion criteria in the scoping review was that studies had to be applicable to cow-calf operations, any studies that had been eliminated because of this criterion were re-examined and included in the present study if appropriate as we did not use this exclusion criterion in the present study.

2.4. Screening and data extraction

Two reviewers screened articles at the title/abstract and full text levels using Covidence online software (Veritas Health Innovation, Melbourne, Australia). The questions asked to pass screening at the title/abstract and full text levels were: Is the full text available in English? Has the study been published in 1950 or later? Is the study an observational or experimental study? Does the study compare vaccination regimes for the prevention of ETEC diarrhea in calves 2 weeks old or younger? Is there a concurrent comparison group? Does the study diagnose the pathogen in diarrheic calves based on lab testing if it is not a challenge trial? Is the study population calves 2 weeks old or younger? Does the study report one of the outcomes diarrhea or death to evaluate the effectiveness or efficacy of the vaccine? Is the study published in a peer-reviewed journal or conference proceedings >500 words or as part of the USDA product summaries for biologics (United States Department of Agriculture, 2024)?

Single data extraction and verification was used, where one reviewer extracted the data and another reviewer verified accuracy of the extracted data. The following items were extracted and recorded in a Microsoft Excel (Redmond, WA) spreadsheet if available from each publication: First author name and year of publication, year when study was conducted, country where study was conducted, production system (beef, dairy, or laboratory), age of calves at enrollment and follow-up time, breed and sex of calves enrolled, housing type during the trial period, publication type (journal, conference abstract, or Center for Veterinary Biologics USDA trial), field or challenge exposure trial, how a diagnosis of ETEC diarrhea was made, how often calves were assessed for disease, how long the follow-up period was, vaccine type (killed, recombinant, or subunit), if a combination vaccine was used, what other pathogens were included, vaccine timing during pregnancy, route of administration, how many doses were delivered, what the vaccine adjuvant was, what the vaccine strain was, whether it was an experimental or commercial vaccine, the control type (placebo, adjuvant, or no vaccine), whether a challenge was given, challenge strain type, challenge dose, if the challenge was homologous or heterologous to the vaccine strain, whether randomization, blinding, or a funding source were mentioned. If multiple trials were conducted as part of the study, we evaluated each trial separately. When multiple treatment groups were compared to only one control group, treatment groups were combined to create a single pairwise comparison to avoid double counting of control group animals in meta-analysis. If there was one control group for multiple vaccine groups, and both heterologous and homologous challenges were included, only heterologous challenge group results were used in data analysis as we deemed heterologous challenge as a more rigorous test of a vaccine. Any data points that did not meet our inclusion criteria were eliminated from the analysis. Some studies had multiple trials, where not all trials met all inclusion criteria and those were therefore eliminated from analysis. For details on decisions for individual trials see supplementary file S2.

2.5. Risk of bias assessment

Risk of bias was assessed using SYRCLE’s risk of bias tool (Hooijmans et al., 2014). Bias domains assessed were generation of random allocation sequence, similarity of groups at baseline, concealment of group allocation, random housing of animals during the experiment, blinding of caregivers and/or investigators to interventions animals received, blinding of outcome assessors, addressing incomplete outcome data, selective outcome reporting, other problems (influence of funders, animals added to replace drop-outs, unit of analysis errors).

2.6. Outcomes

Outcomes in our study were dichotomous, i.e. presence or absence of diarrhea or death in study calves. The authors’ definition for a diarrhea diagnosis was used including various scoring systems or isolation of the target pathogen through culture or other laboratory methods. Risk ratios were calculated with 95 % confidence intervals from the data reported in studies.

2.7. Statistical analysis

Statistical analyses were performed using R Studio (version 2025.05.1, Posit Software, Boston, MA), package ‘meta’ (Balduzzi et al., 2019). Studies with no outcomes in either trial arm were excluded from analysis because a risk ratio cannot be calculated. Heterogeneity in fixed effects models was assessed using I² and a p-value of < 0.1 for evidence of heterogeneity. If significant heterogeneity was present, a cumulative meta-analysis was performed to investigate year of publication as the source of heterogeneity. Contour-enhanced funnel plots were visually inspected to assess the data for the presence of publication bias. Contour-enhanced funnel plots are centered at 1.0, the value under the null hypothesis of no effect. Several levels of statistical significance of the trials are indicated by shaded regions (Viechtbauer, 2010). Trim and Fill method was used as a sensitivity analysis to assess the impact of publication bias (Duval & Tweedie, 2000).

3. Results

3.1. Trial selection

There were 3921 unique citations identified during the initial literature search that mentioned calf diarrhea. Three studies were found as part of the Center for Veterinary Biologics website and one study was found through a SCOPUS search. After eliminating 244 duplicate articles 3677 records were then screened by title and abstract. During this round 3609 records were eliminated for not evaluating vaccine regimes for the prevention of ETEC, one was eliminated since it was published before 1950, and six for not having a concurrent comparison group leaving 61 articles for full text review. In the full text review, eight articles were eliminated for not evaluating vaccine regimens for prevention of ETEC in calves 2 weeks old or younger, six for not having a concurrent comparison group, five field trials had no laboratory diagnosis, eight had a study population older than 2 weeks, and four did not report outcomes death or diarrhea. In two articles, the challenge strain was not defined, making it impossible to categorize the challenge strain as homologous or heterologous to the vaccine strain, four were not published in a peer-reviewed journal or conference proceedings with at least 500 words or as part of the USDA product summaries for biologics, one full text was not available, and one was a case report. A total of 39 articles were eliminated in the full text review leaving 22 articles eligible to be included in the study (Fig. 1). Table 1 is a description of all studies included ordered by year of publication. Study types fell into three main categories: field trials, homologous challenge trials, and heterologous challenge trials. As results could be expected to differ between study types due to various degrees of pathogenic challenge during the trials, we decided that a combined meta-analysis was not appropriate. We therefore performed sub-group analyses for each study type and interpreted the results for each sub-group separately.

Fig. 1.

PRISMA flow diagram of citations found in the literature on vaccination of dams for prevention of calf diarrhea or death due to Enterotoxigenic E. coli in calves from 1950 – 2025.

Table 1.

Publications included in a meta-analysis on the prevention of calf diarrhea due to enterotoxigenic E. coli through dam vaccination published between 1950 and 2025.

Study (First author, publication year)	Total study size	Control type (placebo, no vaccine)	Study type (field or challenge trial)	Calf age from enrollment to end of study	Outcomes
(Gay et al., 1964a)	8 calves	No vaccine	Challenge	1–12 days	Diarrhea
(Gay et al., 1964b)	216 calves	No vaccine	Field trial	≤ 14 days	Diarrhea and Death
(Myers et al., 1973)	35 calves	Placebo	Challenge	0–4 days	Diarrhea
(Newman et al., 1973)	40 calves	Placebo	Challenge	0–2 days	Diarrhea
(Varga & Farid, 1976)	70 calves	No vaccine	Field trial	0–12 days	Death
(Contrepois et al., 1978)	10 calves	No vaccine	Challenge	1–3 days	Diarrhea and Death
(Acres, Isaacson, Babiuk et al., 1979)	34 calves	No vaccine	Challenge	0–10 days	Diarrhea and Death
(Acres, Isaacson, Khachatourians et al., 1979)	43 calves	No vaccine	Challenge	0–10 days	Diarrhea and Death
(Bagley & Call, 1979)	40 calves	Placebo	Challenge	1–4 days	Diarrhea and Death
(Myers, 1979)	62 calves	Placebo	Challenge	1–5 days	Diarrhea
(Acosta-Martinez et al., 1980)	8 calves	No vaccine	Challenge	1–8 days	Diarrhea and Death
(Kornitzer et al., 1980)	55 calves	No vaccine	Challenge	0–8 days	Diarrhea and Death
(Myers, 1980)	82 calves	Placebo	Challenge	0–3 days	Diarrhea
(Nagy, 1980)	22 calves	Placebo	Challenge	0–6 days	Diarrhea and Death
(Snodgrass et al., 1982)	24 calves	No vaccine	Challenge	0–6 days	Diarrhea
(Contrepois & Girardeau, 1985)	12 calves	No vaccine	Challenge	1–5 days	Diarrhea and Death
(Sihvonen & Miettinen, 1985)	200 calves	No vaccine	Field trial	0–15 days	Diarrhea
(Valente et al., 1988)	49 calves	Placebo	Challenge	0–7 days	Diarrhea and Death
(Ohashi et al., 1990)	16 calves	No vaccine	Challenge	0–7 days	Diarrhea and Death
(Jayappa et al., 2008)	29 calves	Placebo	Challenge	1–10 days	Diarrhea and Death
(United States Department of, 2024)	31 calves	Not Mentioned	Challenge	0–10 days	Diarrhea and Death
(Yarnall et al., 2024)	15 calves	Placebo	Challenge	0–14 days	Diarrhea and Death

3.2. Risk of bias assessment

Only one study mentioned how the allocation sequence was generated. Comparing baseline characteristics of study groups, concealment of allocation, blinding of caregivers and/or investigators or those assessing the outcome was almost completely absent from studies. Four articles mentioned the funding source for their studies, and all funders were the manufacturers of the vaccines being tested. None of these studies contained wording that assured that funders had no influence on study design or the conduct of the study. Incomplete or selective outcome reporting, on the other hand, was rare or absent from studies (Fig. 2). Bias in the evaluated studies cannot be ruled out based on our criteria.

Fig. 2.

Bias assessment of articles included in a meta-analysis on vaccination of dams for prevention of calf diarrhea or death due to Enterotoxigenic E. coli in calves from 1950 – 2025.

3.3. Field trials

Three field studies were analyzed, representing eleven trials for the outcomes diarrhea and/or death from ETEC (Table 2). The vaccines tested in field trials were experimental vaccines except for the study by Sihvonen and Miettinen, which tested the commercial vaccine Coligen (Fort Dodge, USA)^, and did not mention the vaccine type or adjuvant used. The oil-adjuvanted vaccine in the study by Gay et al. was administered in a single dose in dams, while vaccines in other field trial studies were given twice. Vaccines used in field trials were either formalin or heat inactivated formulations.

Table 2.

Field trials included in a meta-analysis on the prevention of calf diarrhea or death due to enterotoxigenic E. coli through dam vaccination published between 1950 and 2025.

Study (First Author Name, Year)	Trial Description	Vaccine Strain	Outcomes	Vaccine Type; Route of Administration	Experimental or Commercial Vaccine
Gay, 1964a	Herd 1, vaccine strain 1 N = 24	O26:K60	Diarrhea and death	Formalin-killed oil-adjuvant; intramuscular	Experimental
Gay, 1964a	Herd 1, vaccine strain 2 N = 22	OVC:2984	Diarrhea and death	Formalin-killed oil-adjuvant; intramuscular	Experimental
Gay, 1964a	Herd 2, vaccine strain 1 N = 24	O115KPs3061	Diarrhea and death	Formalin-killed oil-adjuvant; intramuscular	Experimental
Gay, 1964a	Herd 2, vaccine strain 2 N = 14	OVC2995	Diarrhea and death	Formalin-killed oil-adjuvant; intramuscular	Experimental
Gay, 1964a	Herd 3 N = 40	O101:KRVC118	Diarrhea and death	Formalin-killed oil-adjuvant; intramuscular	Experimental
Gay, 1964a	Herd 4 N = 37	O26:K60	Diarrhea and death	Formalin-killed oil-adjuvant; intramuscular	Experimental
Gay, 1964a	Herd 5 N = 32	OVC:250	Diarrhea and death	Formalin-killed oil-adjuvant; intramuscular	Experimental
Gay, 1964a	Herd 6 N = 13	O8:K? and O15:K? O9:K(A) Ps274	Diarrhea	Formalin-killed oil-adjuvant; intramuscular	Experimental
Varga, 1976	Herd 1 N = 70	O101:K30(A)	Death	Heat-inactivated Freund adjuvanted; intramuscular	Experimental
Varga, 1976	Herd 2 N = 54	O8:K28, O101:K30, O78:K80(B)	Death	Heat-inactivated Freund adjuvanted, intramuscular	Experimental
Sihvonen, 1985	1–2 day old calves, N = 101	K99, K35, K30, K85	Diarrhea	Unknown; intramuscular	Commercial, Coligen

In the trials described by Gay et al., 1964a, vaccine strains were matched to strains previously isolated from calves affected by diarrhea in these herds. Authors describe that herds 2 and 4 had a marked drop in diarrhea incidence following vaccination in both vaccinated and control animals and thus contributed little information. Where losses occurred, they were associated with E. coli strains antigenically different from the vaccine strain. In the vaccine trial in the Sihvonen & Miettinen, 1985 study, a commercial vaccine containing unknown serotypes, but characterized as K99 (F5 fimbrial antigen), and capsular polysaccharides K30, K35, and K85 was used and no differences in diarrhea incidence was found between calves from vaccinated and control dams. On the other hand, in the two trials described in Varga & Farid, 1976, only 1/98 calves in the vaccine groups versus 5/26 in the control groups died of ETEC. The authors used a monovalent vaccine with strain O101:K30, which is F41 positive, and a trivalent vaccine with O101:K30, O8:K28, which is F5 and/or F41 positive, and O78:K80, which is known as a cause of calf septicemia. Only the outcome death was evaluated in this study and strain O101:K30 was found in the intestines of calves that died in the control group for the monovalent vaccine trial while O78:K80 was found in control calves for the trivalent vaccine trial.

Taken together, the presented field trial data failed to provide enough evidence that vaccinating dams against E. coli decreased the rate of death or diarrhea within the study populations (Fig. 3). Fixed effects models for the outcomes diarrhea or death showed no evidence of heterogeneity (p = 0.89 for outcome diarrhea and 0.29 for outcome death). Neither model showed a significant effect of dam vaccination on either diarrhea or death in calves (RR 0.89 (95 % CI 0.57 - 1.40) and RR 0.51 (95 % CI 0.22 −1.23) for diarrhea and death, respectively. The funnel plot for the outcome diarrhea showed some asymmetry (Fig. 5A), an indication for publication bias with few smaller studies showing favorable results, while no such asymmetry was found for the outcome death (Fig. 5B). The Trim and Fill meta-analysis for the outcome diarrhea resulted in a RR of 0.92 (95 % CI 0.59 – 1.44), while for the outcome death the RR was identical to the initial model as no studies were filled in (see Figures 6 – 9, supplementary file, S3 file).

Fig. 3.

Forest plots of field trials in a meta-analysis evaluating the efficacy of vaccinating dams with an E. coli vaccine for the prevention of diarrhea (A) death (B) in calves less than 2 weeks old. Trials are from studies published between 1950 and 2025 where a laboratory diagnosis of enterotoxigenic E. coli was made in affected calves.

Fig. 5.

Contour funnel plots of studies evaluating the efficacy of vaccinating dams with an E. coli vaccine for the prevention of diarrhea (A) or death (B) in field trials in calves less than 2 weeks old or for the prevention of death in trials using heterologous challenge (C) or homologous challenge (D) to the vaccine strain. Trials are from studies published between 1950 and 2025. Funnels are centered around 1.00, the value under the null hypothesis of no effect. Shaded regions correspond to p-values > 0.1 (white), 0.1 > p > 0.05 (dark grey), 0.05 > p > 0.01 (medium grey) and < 0.01 (area outside of funnel, light grey). The vertical dashed line represents the estimated treatment fixed effect. Diagonal dashed lines represent the 95 % confidence interval for the treatment effect. The dotted vertical line in C) represents the random effects estimate.

3.4. Heterologous challenge trials

In fourteen studies representing fifteen trials, a mix of experimental and commercial vaccines were tested using heterologous challenge exposure (Table 3). Five trials included commercial vaccines. Vaccines consisted of mono or multivalent strains. Strain B44 was used in six of the trials as the heterologous challenge strain. B44 is a reference strain originally isolated from a diarrheic neonatal calf in the UK and expresses both F5 and F41 fimbriae (Karkhanis & Bhogal, 1986). Challenge doses ranged between 1 × 10¹⁰ to 4 × 10¹² colony forming units (CFU). Two trials did not mention the challenge dose. The vaccines were described mostly as formalin-killed whole cell bacterins, or “E. coli pili antigen” or “K99 antigen”. In the trials evaluated, when calves from vaccinated dams were given a heterologous challenge there was evidence of a decrease in the rate of diarrhea and death caused by E. coli. However, there was evidence of heterogeneity (P < 0.01) among trials for the outcome diarrhea. After assessing a cumulative meta-analysis model, there was no indication of a change in outcomes at any particular study (see Figure 9, supplementary file S3). Therefore, meta-analysis for the outcome diarrhea was no longer pursued for heterologous challenge trials as no source of heterogeneity was apparent. For the outcome death, the P-value for heterogeneity was above our predetermined cut-point of 0.1 (P = 0.16) and meta-analysis resulted in a RR of 0.25 (95 % CI 0.18 - 0.35) (Fig. 4A). The funnel plot for the outcome death showed evidence for publication bias with fewer small studies showing a risk ratio above the model estimate (Fig. 5C). Trim and Fill analysis for the outcome death resulted in a RR of 0.32 (95 % CI 0.23 – 0.43) after filling in five studies, a slight increase from the estimate of 0.25 prior to Trim and Fill but still far from a RR of 1.0, which would indicate no effect (see Figures 11 −12, supplementary file S3).

Table 3.

Challenge trials using heterologous challenge strains included in a meta-analysis on the prevention of calf diarrhea or death due to enterotoxigenic E. coli through dam vaccination published between 1950 and 2025.

Study (First Author Name, Year)	Trial Description, N for trial	Vaccine Strain(s)	Challenge Strain	Challenge Dose (CFU)	Outcome	Vaccine Type; Route of Administration	Experimental or Commercial Vaccine
Gay, 1964b	Trial 4, Ps125 M vaccine strain N = 3	Ps125M(O26:K60(B))	RVC 1787 O137:K79:H41	Not mentioned	Diarrhea	Formalin-killed whole cell bacterin; intramuscular	Experimental
Acres, 1979a	Purified K99,	Purified K99: Strain 1474 (K12 and K99+)	B44	6 × 1011	Diarrhea and Death	Purified E. coli pili; subcutaneous	Experimental
	Multiple strain bacterin N = 25	Multiple strain bacterin: O9:K25, O8:K85, O9:K35, O29:K?, O101:K28, O101:K30	B44	6 × 1011	Diarrhea and Death	Formalin-killed whole cell bacterin (MSB); subcutaneous	Commercial (Multiple strain bacterin, Fort Dodge, IA)
Acres, 1979b	Purified K99	Purified K99 antigen	B44	1 × 1011	Diarrhea and Death	Purified K99 antigen; subcutaneous	Experimental
	K99+ ALEC	anucleated live E. coli (K99+ ALEC)	B44	1 × 1011	Diarrhea and Death	K99-positive anucleated live E. coli; subcutaneous	Experimental
	Multiple strain bacterin N = 34	Multiple strain bacterin: (O8:KK25, O8LK85, O9:K35, O20:K?, O101:K28, O101:K30)	B44	1 × 1011	Diarrhea and Death	Formalin-killed whole cell bacterin with six strains; subcutaneous	Commercial (Fort Dodge, IA)
Myers, 1979	B41 vaccine strain N = 19	B41 (O101:K-:K99)	O9:K35:K99	1 × 1010 - 3 × 1010	Diarrhea	Formalin-killed whole cell bacterin with 1 strain; intramuscular	Experimental
Kornitzer, 1980	Challenge strain B44, N = 13	Strain 168 (O101:K30:H9:K99)	B44 (O9:K30:H-:K99, Ent+)	2.5–5.3 × 1011	Diarrhea and Death	K99 antigen; subcutaneous	Experimental
Kornitzer, 1980	Challenge strain B42, N = 13	Strain 168 (O101:K30:H9:K99)	B42 (O9:K35:H-:K99, Ent+)	3.5–5.3 × 1011	Diarrhea and Death	K99 antigen; subcutaneous	Experimental
Myers, 1980	Experiment 1, K99 bacterin, N = 41	B41, O101:K-:K99	O9:K35:K99 or O8:K85:K99	5 × 1010	Diarrhea and Death	Killed; injection route not specified	Experimental
Nagy, 1980	K99 exctract, N = 22	K-12, K99	K-12, K99	1.6 × 1011	Diarrhea and Death	K99 antigen; subcutaneous	Experimental
Snodgrass, 1982	0.5 ml oil-adjuanted, 2 ml oil-adjuvanted, 4 ml aluminum-hydroxide adjuvanted vaccine, N = 24	ETEC 1 (O101:K99)	B44 (ETEC 2) (O9:K30(B):K99)	4.1 × 1010	Diarrhea	Crude K99 extract; intramuscular	Experimental
Contrepois, 1985	Colostrum types A, B, A + B N = 12	O101:K99:H- (B41)	B41	5 × 1011 - 1 × 1011	Diarrhea and Death	Purified K99 antigen; subcutaneous	Experimental
Valente, 1988	B41 vaccine strain N = 49	B41 (O101:K-:H-, K99, F41)	B44 (O9:K30:H-, K99, F41)	4 × 1010	Diarrhea and Death	E. coli pili antigen; subcutaneous	Experimental
Ohashi, 1990	Heterologous challenge strain N-5 N = 8	NADC 1471 (O101, K-, K99 F41)	N-5 (O9, K35, K99)	1 × 1010 −1 x 1011	Diarrhea and Death	Formalin-killed whole cell bacterin; administration route not specified	Commercial (Norden Laboratories, NE)
Jayappa, 2008	Guardian commercial vaccine N = 29	K99 (Strain not mentioned, but identified as heterologous to challenge strain)	E. coli B44 [O9:K30:K99]	4 × 1012	Diarrhea and Death	K99 subunit antigen; subcutaneous	Commercial (Guardian, Schering-Plough Animal Health, NJ)
United States Department of Agriculture, 2024	Bovilis N = 29	K99	Strain not mentioned, but identified as heterologous to vaccine strain	Not mentioned	Diarrhea and Death	Not described; subcutaneous	Commercial (Bovilis, Merck Animal Health, NJ)
Yarnall, 2024	Challenge Strain O101:K99, F41 N = 15	O8:K99	EC 8044 [O101:K99, F41]	10 ml orally 1010 CFU/ml	Diarrhea and Death	No specifics mentioned	Experimental

Fig. 4.

Forest plots of studies using a heterologous (A) or homologous (B) challenge to the vaccine strain in a meta-analysis evaluating the efficacy of vaccinating dams with an E. coli vaccine for the prevention of death in calves less than 2 weeks old. Trials are from studies published between 1950 and 2025.

3.5. Homologous challenge trials

Eight trials where calves from vaccinated dams were given a challenge homologous to the vaccine strain, were analyzed. Four of the trials used strain B44 for both vaccine and challenge (Table 4). Only one trial tested a commercial vaccine. Most vaccines were formalin-killed whole cell bacterins, one vaccine was heat killed, while one trial used concentrated crude toxin and live attenuated bacteria in their vaccine. There was a decreased risk of death and diarrhea from E. coli. While the model for the outcome diarrhea showed evidence of heterogeneity (P < 0.01), this was not the case of the outcome death (P = 0.94). Cumulative meta-analysis showed the results from trial Gay, 1964b Trial 5 had both a large confidence interval as well as a drastically different RR estimate (see Figure 13, supplementary file S3). Meta analysis was therefore repeated without this study. However, heterogeneity persisted with I² = 62 % (P = 0.01) (see Figure 14, supplementary file S3). Similar to the heterologous challenge trials, meta-analysis was no longer pursued for the outcome diarrhea after no source of heterogeneity could be detected. The RR in the fixed effects model for the outcome death was 0.09 (95 % CI 0.03 - 0.31) (Fig. 4B). The corresponding funnel plot does not show clear evidence for publication bias (Fig. 5D). Trim and Fill analysis for outcome death resulted in a RR which is smaller than for the available studies, 0.08 (95 % CI 0.03 – 0.26), as the software filled in one study with a RR below 0.09 (Figures 15 - 16, supplementary file S3).

Table 4.

Challenge trials using homologous challenge strains included in a meta-analysis on the prevention of calf diarrhea due to enterotoxigenic E. coli through dam vaccination published between 1950 and 2025.

Study (First Author Name, Year)	Trial Description	Vaccine Strain	Challenge Strain	Challenge Dose (CFU)	Outcome	Vaccine Type	Experimental or Commercial Vaccine
Gay, 1964b	Trial 5 N = 5	RVC1787 (O137:K79:H41)	RVC1787 (O137:K79:H41)	Only volume mentioned (5 ml)	Diarrhea	Formalin-killed whole cell bacterin; intramuscular	Experimental
Myers, 1973	B44 strain vaccine N = 35	B44	B44	3 × 1011–1.5 × 1012	Diarrhea	Concentrated crude toxin; subcutaneous and intramamary	Experimental
						Formalin-killed whole cell; subcutaneous and intramamary	Experimental
						Live nonattenuated bacteria; subcutaneous and intramamary	Experimental
Newman, 1973	2 doses in current year	B44	B44	3 × 103	Diarrhea	Formalin-killed whole-cell bacterin; subcutaneous	Experimental
	1 dose in current year and 2 doses in previous year					Formalin-killed whole-cell bacterin; subcutaneous and intramammary
	2 doses in previous year N = 40					No vaccine
Contrepois, 1978	B41 strain vaccine N = 10	O101:K99:H- (B41)	B41	5 × 1010 - 1 × 1011	Diarrhea and Death	purified K99 antigen; subcutaneous	Experimental
Bagley, 1979	SQ and IM vaccines N = 40	B44	B44	3 × 1011 (10 ml of 3 × 1010 culture)	Diarrhea and Death	Formalin-killed whole cell bacterin; subcutaneous or intramuscular	Experimental
Acosta-Martinez, 1980	Group 1a: homologous challenge, N = 8	B44 (O9:K30:K99)	B44	1.6 × 1011 – 60 × 1011	Diarrhea and Death	Heat-killed bacterin; intramuscular	Experimental
Kornitzer, 1980	Challenge strain 168, N = 29	Strain 168 (O101:K30:H9:K99)	Strain 168 (O101:K30:H9:K99)	2.5–5.3 × 1011	Diarrhea and Death	Unknown	Experimental
Ohashi, 1990	Homologous challenge strain 880, N = 8	NADC 1471 (O101, K-, K99 F41)	Strain 880 (O101, K?, K99)	1 × 1010 -1 × 1011	Diarrhea and Death	Formalin-killed whole cell bacterin; administration route not specified	Commercial (Norden Laboratories, NE)

4. Discussion

Although vaccines for E. coli have been developed and tested for decades, no previous meta-analyses have been published on this topic. The present study, therefore, presents the available data in an organized manner for the first time. In the present study, the available evidence for vaccination of dams for E. coli showed no effectiveness in preventing death or diarrhea in field trials. However, in both heterologous and homologous challenge trails vaccines helped in preventing death of calves. The sparsity of field trials that met our inclusion criteria, combined with the quality and statistical power in the available studies may be the reason results from challenge trials could not be replicated in the field.

We eliminated several field trials from analysis because they did not establish a laboratory diagnosis to confirm that ETEC was the cause of death or diarrhea in study calves. With multiple possible calf diarrhea agents, it is not possible to evaluate the etiology based on clinical signs alone. Updated field trials that include a laboratory diagnosis would be desirable, although we recognize the logistical challenges involved in such studies.

One of the reasons we may not have been able to analyze data for the prevention of diarrhea in both homologous and heterologous challenge trials is that there is no standardized definition of diarrhea. We relied on definitions or diagnoses provided in the respective studies to arrive at the number of cases of diarrhea among study calves. Several studies used a scoring system to come to a diagnosis of diarrhea, however there may have been some variability in how each individual researcher defined a diarrhea diagnosis. While some studies considered any deviation from normal feces to be diarrhea, others may have used a category of “soft feces” but with no clinical sign of depression or dehydration to be normal. Risk of diarrhea could not be evaluated due to significant heterogeneity between studies that could not be explained with the help of a cumulative meta-analysis, meaning that a particular year did not change the estimate.

Visually inspecting the funnel plots for the analyses performed did not reveal gross evidence of publication bias. The Trim and Fill method resulted in similar results that did not alter the main findings. However, it must be recognized that the Trim and Fill method only serves as a sensitivity analysis and does not result in an accurate unbiased estimate. Therefore, undetected publication bias may still be present.

Many older studies do not comply with the requirements laid out in the REFLECT guidelines for clinical trials for livestock and food safety trials (O'Connor, 2010). Over recent years, progress has been achieved in complying with more rigorous study designs in veterinary medicine, however, the mention of randomization of study animals to treatments or blinding of research personnel is almost completely absent in almost all studies included in the analysis. Poor quality of reporting in animal studies is not uncommon. A study of original research published between 1999 and 2005 on live rats, mice, and non-human primates found that 87 % did not use randomization, 86 % did not use blinding, and only 70 % described their statistical methods and used a measure of variability to report results (Kilkenny et al., 2009). Many trials in this study also had very small sample sizes, in particular those that were part of challenge trials, which are resource intensive and may have been limited by available funds. Some studies also had a plethora of individual trials with a few calves using a variety of vaccine or challenge strains which made data extraction and analysis difficult. Studies with only a few calves may have also been underpowered and sample size calculations are also absent from all publications. Statistical power to detect differences is therefore hard to assess as we cannot replicate assumptions made by researchers on the expected incidence and variability.

Another source of potential heterogeneity encountered during meta-analysis for the diarrhea models could have been the large variety of vaccine and challenge strains. However, the models evaluating the outcome of death did not suffer the same heterogeneity problems, which may indicate that the method of diagnosing diarrhea may have been the bigger source of heterogeneity. With very few studies on commercial products available, we did not deem it feasible to perform a commercial vaccine subgroup analysis.

Other meta-analysis articles evaluating veterinary study topics have also experienced a lack of blinding and randomization reporting. In a publication by Burns and O’Connor where vaccine efficacy in preventing pinkeye in beef cattle was the topic, among 123 vaccine trials to evaluate the efficacy of pinkeye vaccines only 15 reported blinding and randomization (Burns & O’Connor, 2008). In a meta-analysis on the vaccine efficacy of preventing urinary shedding of Leptospira in cattle, 11 out of 12 articles were unclear when reporting randomization and 7 out of 12 were unclear in reporting blinding (Sanhueza et al., 2018). Baltzell’s 2013 critical review on the efficacy of whole cell Tritrichomonas foetus vaccine to prevent and treat trichomoniasis in beef cattle, reported that only 1 out of 10 articles used for the review reported using randomization (Baltzell et al., 2013). Lack of blinding and randomization in studies could lead to biased results which could in turn affect accurate reporting. Reporting guidelines for clinical trials, such as REFLECT reporting guidelines for the randomized controlled trials for livestock and food safety are available now so that future reviews are expected to include studies with more rigorous reporting (O'Connor et al., 2010).

It is unclear whether any of the vaccines described in the studies included in this meta-analysis remain commercially available. The companies Fort Dodge, Norden Laboratories, and Schering-Plough—whose products were commonly referenced—no longer exist under those names. Schering-Plough was acquired by Merck Animal Health, and Fort Dodge was acquired by Pfizer and later sold to Boehringer Ingelheim. While licensed veterinary vaccines are listed in the USDA product catalog, details such as antigen strain identity, concentration, and adjuvant formulation are proprietary and not publicly disclosed. Therefore, without access to current product formulations or recent field studies, it remains unclear how findings from older vaccine studies relate to contemporary products.

5. Conclusions

There is a lack of high quality randomized controlled field trials that would allow a reliable assessment of efficacy of ETEC vaccines in preventing calf morbidity and mortality. In addition, most studies available suffer from some risk of bias due to a lack of blinding and randomization. Regardless of the limitations, the present study provides evidence that vaccination of dams with ETEC vaccines during gestation can prevent death in calves from enterotoxigenic E. coli infections.

Funding

This work was supported by a grant from the California Department of Food and Agriculture, agreement number 18–0623–000-SG

Ethical statement

“Dam vaccination for the prevention of neonatal diarrhea caused by enterotoxigenic E. coli in calves-a systematic review and meta-analysis” by Higgs et al.

The research conducted in this manuscript meets the criteria for conduct, reporting, editing and publication of scholarly work in medical journals.

In particular, guidelines on authorship, disclosure of financial support, conflicts of interest, non-author contributions, duplicate submissions and prior publication have been followed.

No animal or human subjects have been involved in the research; therefore, no corresponding approvals of review boards are necessary.

CRediT authorship contribution statement

Chandler Higgs: Writing – original draft, Investigation, Data curation, Conceptualization. Megan Van Noord: Writing – review & editing, Investigation, Data curation. Jefferson Gabriel Carvalho Nagle: Writing – review & editing, Data curation. Jose Pablo Gomez: Data curation. Erik Fausak: Writing – review & editing, Investigation, Data curation. Gabriele Ute Maier: Writing – review & editing, Supervision, Project administration, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Gabriele Maier reports financial support was provided by California Department of Food and Agriculture. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix. Supplementary materials

Supplementary materials

Supplementary materials S1 - 3 associated with this article can be found, in the online version.