MATERIALS AND METHODS

Ethical approval

This study was based exclusively on retrospective reproductive and herd health records routinely collected between 2021 and 2023 from a dairy cooperative in Malang Regency, East Java, Indonesia. No experimental interventions, invasive procedures, direct animal handling, biological sample collection, or manipulation of animals were performed specifically for research purposes. All data were generated during routine veterinary and reproductive management practices. Before analysis, the dataset was anonymized to remove all farmer and animal identifiers, thereby ensuring confidentiality and data privacy. In accordance with institutional and national guidelines governing research based solely on retrospective secondary records without direct animal involvement, formal approval from an animal ethics committee was not required.

Study period and location

The study was conducted using reproductive records collected from a dairy cooperative comprising smallholder dairy farms in Malang Regency, East Java, Indonesia. The study period covered January 2020 to August 2023, encompassing both pre-outbreak and post-outbreak reproductive records. The FMD outbreak occurred between May 24 and August 8, 2022. Initial outbreak confirmation was performed by the Veterinary Diagnostic Laboratory (Wates, Yogyakarta, Indonesia), the national reference laboratory for FMD diagnosis. Because laboratory confirmation was not feasible for every affected animal during the outbreak, subsequent cases were diagnosed clinically by trained local veterinarians based on compatible clinical signs, including oral ulcers, hypersalivation, pyrexia, anorexia, lameness, and ulcerative lesions affecting the extremities.

Each animal maintained within the cooperative possessed an individual identification record linked to its ear tag, enabling complete documentation of artificial insemination (AI) services, reproductive status, pregnancy diagnosis, vaccination history, calving records, and the onset of clinical FMD.

During the national vaccination campaign, clinically healthy cattle received a primary dose of the inactivated FMD vaccine Aphtovaks-E™ (PT Vaksindo Satwa Nusantara, Bogor, Indonesia) in June 2022, followed by a booster dose of Avtogen™ (Biogénesis Bagó, Garín, Buenos Aires, Argentina) in July 2022. The heifers included in this study had no previous history of FMD vaccination; therefore, the June 2022 vaccination represented their first exposure to FMD immunization.

Study design

This retrospective, within-animal, before-and-after observational study evaluated reproductive performance before-and-after natural FMD infection or vaccination, using paired reproductive records from the same animals. The within-animal design enabled each heifer to serve as its own control, thereby minimizing between-animal variability associated with genetics, management, and environmental factors.

Reproductive performance was compared during two consecutive observation periods before-and-after FMD exposure. The same heifer contributed data to both observation periods, allowing direct within-animal comparison of fertility outcomes following natural infection or vaccination.

Study population

A total of 772 HF heifers from 501 smallholder dairy farmers were included in the study. All heifers received their first AI between 15 and 18 months of age.

For post-exposure evaluation, animals were classified into three mutually exclusive exposure groups according to their clinical status and the timing of FMD relative to vaccination:

1. Heifers that developed clinical FMD during the outbreak (Condition 1; n = 464).

2. Heifers that remained clinically healthy during the outbreak but subsequently developed clinical FMD after the first vaccination round (Condition 2; n = 60).

3. Heifers that remained clinically healthy throughout the outbreak and after the first vaccination round (Condition 3; n = 248).

The selection process for the study animals is illustrated in Figure 1, whereas baseline characteristics by exposure condition are presented in Table 1. All 772 heifers contributed paired pre-exposure and post-exposure reproductive records; therefore, no single-period records were included in the final analyses. The study timeline is presented in Figure 2.

Figure 1: STROBE flow diagram showing the selection of Holstein–Friesian heifers included in the within-animal before-and-after analysis.

Figure 2: Schematic timeline illustrating the observation periods for reproductive performance in Holstein-Friesian heifers before-and-after the foot-and-mouth disease (FMD) outbreak in Indonesia. The pre-FMD observation period included heifers first inseminated between January 8, 2020, and November 30, 2021, whereas the post-FMD observation period included those first inseminated between June 6, 2022, and October 31, 2022. Each observation period lasted 180 days from the first service (FS), defined as the first artificial insemination.

Characteristics | Condition 1 | Condition 2 | Condition 3
Exposure definition | Clinical FMD during the outbreak | Clinically healthy during the outbreak but developed clinical FMD after the first vaccination | Clinically healthy during the outbreak and remained free of clinical FMD after the first vaccination
Number of heifers, n | 464 | 60 | 248
Number of smallholder farmers, n | 362 | 50 | 204
Pre-exposure first AI period | January 8, 2020–November 30, 2021 | January 8, 2020–November 30, 2021 | January 8, 2020–November 30, 2021
Post-exposure first AI period | June 6, 2022–August 11, 2023 | June 6, 2022–August 11, 2023 | June 6, 2022–August 11, 2023

Eligibility criteria

HF heifers were eligible for inclusion if they:

1. were registered in the cooperative database;

2. possessed identifiable individual records linked to ear tag information;

3. had complete AI and pregnancy diagnosis records during both pre-exposure and post-exposure observation periods;

4. received their first AI between 15 and 18 months of age; and

5. could be assigned to one of the three predefined exposure conditions.

Heifers were excluded if they lacked complete reproductive records before-and-after FMD infection or vaccination, if the first AI did not occur within the predefined 6-month pre-FMD observation period, or if essential information regarding AI date, pregnancy diagnosis, FMD status, or vaccination history was incomplete. The selection process and reasons for exclusion are presented in Figure 1.

Observation window and time origin

Reproductive performance was evaluated during two 180-day observation periods. During the pre-outbreak period, heifers receiving their first AI between January 8, 2020, and November 30, 2021, were followed for 180 days from the date of first insemination.

During the post-exposure period, follow-up commenced from the first AI performed after the relevant exposure date, defined as the date of clinical FMD onset for clinically affected heifers or the vaccination date for vaccinated clinically healthy heifers. The first AI initiating post-exposure follow-up occurred between June 6, 2022, and August 11, 2023. All inseminations occurring within the same 180-day period were considered part of the corresponding observation window.

Outcome definitions

The primary reproductive outcomes were defined a priori. FSCR was defined as conception following the first AI within a 180-day observation period (yes/no) [32].

PR was defined as achievement of at least one confirmed pregnancy during the same 180-day observation period (yes/no) [14, 33].

S/C was defined as the number of inseminations required to achieve one confirmed pregnancy among heifers that conceived during follow-up [34].

First-service-to-conception interval (FSCI) was defined as the number of days from the first AI to the insemination resulting in confirmed pregnancy. Heifers conceiving at first-service were assigned an FSCI of 0 days.

Pregnancy diagnosis was performed approximately 60 days after AI by trained veterinary personnel using rectal palpation. A confirmed pregnancy was recorded following a positive rectal palpation result in the cooperative reproductive database. Embryonic or fetal loss was inferred when a heifer with a confirmed pregnancy failed to produce a subsequent calving record.

Proxy definition of reproductive record exit

Complete records describing culling, sale, transfer, or mortality were unavailable; therefore, reproductive record exit was evaluated using a predefined proxy outcome.

A heifer was considered to have experienced reproductive record exit if, following the exposure date, she had no additional AI records during the 180-day follow-up period and no confirmed pregnancy record. The date of the final recorded AI was considered the proxy for the reproductive record exit date.

This definition was intended solely to identify the discontinuation of reproductive follow-up and should not be interpreted as confirmed culling due to infertility. Rather, it represents a pragmatic outcome derived from routinely collected AI records in smallholder dairy production systems where complete culling records are unavailable.

Statistical analysis

FSCR and PR were analyzed using mixed-effects logistic regression models. Observation period (before versus after FMD), exposure condition, and their interaction were included as fixed effects, whereas heifer identity was incorporated as a random intercept to account for repeated measurements. Farm-level clustering was evaluated as an additional random effect and retained when supported by model fit. Results are presented as odds ratios (OR) with 95% confidence intervals (CI).

An exploratory post-exposure analysis evaluated FSCR according to the interval between clinical FMD onset or vaccination and the first post-exposure AI. Intervals were categorized as 0–30, 31–60, 61–90, and >90 days, with the >90-day interval serving as the reference category. Pairwise ORs were calculated within each exposure condition, and Fisher's exact test was used to compare each interval with the reference category. When zero conception events occurred, the Haldane–Anscombe correction was applied. Because of sparse observations, particularly in Condition 2, the results of this exploratory analysis were interpreted with caution.

S/C was summarized descriptively among heifers that conceived during follow-up, and no formal statistical comparisons were performed.

FSCI was analyzed as a time-to-event outcome using Kaplan–Meier survival analysis and Cox proportional hazards regression. Time was defined as the interval from first AI to confirmed conception, whereas non-pregnant heifers were right-censored at 180 days. For computational purposes, heifers conceiving at first-service were assigned an FSCI of 1 day. Cox models included the observation period, exposure condition, and their interaction, with robust standard errors clustered at the heifer level. Results are presented as hazard ratios (HR) with 95% CI.

Reproductive record exit was evaluated only during the post-FMD period using Kaplan–Meier survival analysis and Cox proportional hazards regression. Because the proportional hazards assumption was violated according to Schoenfeld residuals, a piecewise Cox proportional hazards model was fitted by dividing follow-up into 0–90-day and 91–180-day intervals and incorporating condition-by-time interactions.

All statistical analyses were two-sided, and statistical significance was defined as p < 0.05. Analyses were performed using R version 4.4.1 with the lme4**, survival, dplyr, tidyr, **ggplot2, and survminer packages (R Foundation for Statistical Computing, Vienna, Austria).