Association between previous days open and production performance in older Holstein cows under Swiss grassland-based...

The objectives of the present study were to investigate the association between previous days open (PDO) and subsequent lactation performance in the older Holstein cows managed under Swiss grassland-based dairy production systems. We hypothesized that excessively extended PDO in older Holstein cows would be negatively associated with subsequent 305-d lactation performance. A retrospective cohort study was conducted including 879,978 cow-lactations from lactations 3 (L3; n = 412,898), 4 (L4; n = 284,391), and 5 (L5; n = 182,689) across 21,335 Swiss dairy farms. Each cow-lactation contributed an average (±SD) of 8.10 ± 0.96 (L3), 8.05 ± 0.98 (L4), and 8.03 ± 0.99 (L5) monthly test-day records, collected between 1995 and 2023 within 305-d lactation, yielding a total of 7.10 million test-day records (L3 = 3.34; L4 = 2.29; L5 = 1.47 million). Previous days open, calculated from the calving interval adjusted for gestation length, averaged 120 ± 59 d (L3), 120 ± 58 d (L4), and 121 ± 59 d (L5), with a range of 40 to 360 d. Data on daily milk yield and milk composition and SCC were collected. Yields of daily fat, true protein, lactose, and ECM were calculated. Generalized linear mixed-effects models included fixed effects of linear and quadratic covariates of PDO, lactation, a natural cubic spline for DIM, linear covariate of year, and all possible 2-way interactions, and random effects of farm nested within monthly test-day record and cow. Model-predicted estimates are presented at representative PDO values of 80, 160, and 240 d, corresponding to conventional, moderately extended, and excessively extended reproductive intervals, whereas all statistical analyses treated PDO as a continuous variable across its observed range (40 to 360 d). An interaction between the quadratic covariate of PDO and lactation number was observed for yields of milk, ECM, protein, lactose, and SCS, indicating that the curvilinear response to PDO differed among lactations. Based on model-predicted estimates at representative PDO values, increasing PDO from 80 to 240 d was associated with a change in daily milk yield by +1.70, +1.10, and +0.60 kg/d; ECM yield by +2.30, +1.60, and +1.20 kg/d; true protein yield of +0.02, +0.01, and -0.01 kg/d; and lactose yield of +0.08, +0.05, and +0.03 kg/d for L3, L4, and L5, respectively. Across the representative PDO range evaluated (80 to 240 d), predicted SCS increased from 1.64 to 1.82, 1.85 to 2.07, and 2.02 to 2.27 for L3, L4, and L5, respectively. Quadratic association was observed between PDO and fat yield with increasing PDO from 80 to 240 d associated with changes of +0.12, +0.10, and +0.08 kg/d for fat yield in L3, L4, and L5, respectively. These observed quadratic associations were lactation and trait-specific, with the PDO values associated with maximum predicted daily milk, ECM, and individual milk component yields differing among L. Overall, the results suggest that both excessively short and excessively long reproductive intervals may compromise different aspects of subsequent lactation performance under the conditions evaluated, with the optimal PDO varying according to production trait and lactation number.