Age at first calving in dairy cows: which months do you aim for to maximise productivity?

PICO question 
In dairy cattle, which months should producers target age at first calving in order to maximise milk yield, minimize risk of non-voluntary culling and optimize fertility? 
  
Clinical bottom line 
Category of research question 
Risk 
The number and type of study designs reviewed 
Seventeen papers were critically reviewed: 15 sets of case series, one review of case series and cohort studies and one randomised control trial, summarising over 2.4 million individual cow records 
Strength of evidence 
Strong 
Outcomes reported 
An optimum range of age at first calving (AFC) on dairy farms appears to be 22–25 months inclusive. Lower or higher than this figure can bring lower first lactation 305 day and lifetime milk yields, lower fertility and lower chances of surviving to a second lactation. Achieving an AFC of 22–25 months can bring the highest economic return to dairies 
Conclusion 
Age at first calving is a useful and key performance parameter to measure in dairy cattle. Achieving a range of 22–25 months at first calving can help to optimise both long term milk yield, fertility and longevity within the herd 
  
How to apply this evidence in practice 
The application of evidence into practice should take into account multiple factors, not limited to: individual clinical expertise, patient’s circumstances and owners’ values, country, location or clinic where you work, the individual case in front of you, the availability of therapies and resources. 
Knowledge Summaries are a resource to help reinforce or inform decision making. They do not override the responsibility or judgement of the practitioner to do what is best for the animal in their care. 
  



Milk Yield
Milk yield (MY) has been measured as the volume of milk harvested from the cow in the 305 day lactation following her first calving (taken by addition of test-day results, daily milk weights or predicted 305 day yield in papers where yield was recorded over less than 305 days). Prediction of 305 day milk equivalent is a common parameter from dairy software recording systems that use equations linked to prediction of the lactation curve. Cows at certain days in milk (DIM) can have known monthly test day results to base a prediction figure from and calculate their predicted yield to the end of that lactation. Another parameter reported has been in MY from the cow over her total productive lifetime (either in total milk volume or average daily yield).
In terms of volume, milk losses and gains can be regarded from the perspective of pre-22 months; 22-25 months and post 25 months AFC. At an AFC of less than 22 months, 590-800 kg losses in the first 305 day lactation have been reported by (Elahi Torshizi, 2016) and (Pirlo, 2011). Ettema (2004) included pre-23 months in his study and reports a 320 kg loss in the first lactation: the lower volume lost is most likely because the cows at 22-23 months were included in the calculation. Only one study reports higher yields in 18-23 months, quoting 4.5% more milk from the first 305 day lactation (Banos, 2007). Cows with an AFC of greater than 25 months are also reported to have lower milk yields. Elahi Torshizi (2016) and Pirlo (2011) report a 170-600 kg loss in milk above 26 months. Berry (2009) reported that first lactation 305 day yield decreases by 55.5 kg less per month, increasing from 22 months AFC to 38 months. Cows calving between 22-25 months have been found to produce 2.1-2.4 kg/day more milk than their counterparts calving outside this window of time (Storli, 2017 andEastham, 2018).
In summary, there appear to be lower milk yields of between 170 and 600 kg in the first lactation in cows with low AFCs (18-21 months) and 590-800 kg less yield in cows with high AFCs (>26 months).

Fertility
There is less clarity over fertility effects of varying age at first calving, mainly due to the parameters that researchers have measured on performance of reproduction. This is discussed in the appraisal section below. First service conception rate was reported by Ettema (2004) to be highest in AFC of <23, then 23-25, then >25 months (75%, 64%, 45%). Conversely, days open and calving interval have been reported as improving in older AFC groups (>26 months) compared to AFCs of <26 months (Eastham, 2018) and (Krpalkova, 2014). Banos, (2007) suggested that fertility would be compromised in younger AFC cattle by reporting 7% more inseminations per pregnancy and 7.5% higher return rate in AFC between 18-23 months. In summary, although more papers suggest higher fertility rates in older AFC cattle, one key paper reflects that aiming for lower AFC does not compromise first service conception rate targets (Ettema, 2004). Longevity There has been no significant, direct effect of AFC on the survival length of cattle reported Nilforooshan (2004 andWathes (2014). However, it must be noted that longevity of cattle is strongly linked to productivity and less productive cattle are more likely to be culled. Therefore, links can be made between AFC and survivability: Eastham (2018) suggests that cows calving between 22-26 months old are more likely to survive to calve a second time but this is not a reflection of the AFC itself but rather the management of transition and early lactation as a whole.

Profitability
Whether or not AFC affects the profitability of a dairy is a difficult and highly variable parameter to calculate, as so many factors in the management of both heifer rearing and lactation affect dairy profit. However, five of the reviewed papers have attempted to associate AFC with profitability. The longer it takes for a heifer to enter the milking herd, the more feed and management costs are involved in rearing her. Ettema (2004) observed that an AFC of 22-24 months was US$98-138 preferable per heifer than other AFCs. Pirlo (2011) calculated that an AFC of 22-26 months improved income per heifer by US$24-41 over other AFCs and Changee (2013) reported that an AFC of 22.5 months gave up to US$727 higher lifetime returns compared to an AFC of 32 months. Krpalkova (2014) suggests that between 24-26 months of AFC returns the highest profitability but does not state the amount: the statement is based on higher milk returns and fertility parameters. Wathes (2014) simply reviews the above papers. In summary, it seems that an AFC of between 22-26 months produces the highest returns compared to lower or higher AFCs.  Every 1 kg of 6 week ADWG was associated with 544 kg more first-lactation 305 day milk and every 1 kg of 8 week ADWG yielded 579 kg more first-lactation 305 day milk but there was great variation around the mean, so BW was more associated with first lactation yield than ADWG. This is true, due to BW being more related to birth weight and this is strongly correlated to first lactation yield  Starter pellet intake had a significant effect on first lactation yields, with each 1 kg increase in intake resulting in 8.21 kg more 305 day milk (but variations were high). Milk replacer intake had no effect on first lactation yield  Season of birth: summer calves had a slightly greater first lactation yield, probably due to photoperiod on intakes Limitations: Small number of farms in the sample population, so other herd management factors could have influenced differences between the three herds.

Acronyms
There was no association with AFC and 305 day yield but the variation between the three farms was small (715, 702 and 725 days).  AFC of <24 months was associated with 727 kg more lifetime yield (lifetime value of energy corrected milk 25,184 kg) than 24-31 months (lifetime value of energy corrected milk 24,387 kg) and 4,401 kg more lifetime yield than >31 months (lifetime value of energy corrected milk 20,783 kg)  There was no significant difference between daily milk yield (over lifetime: lifetime range was 5.4 to 5.9 years, these being least square mean values) and AFC up to 31 months but these were 1 kg/day higher than over 31 months  There was no significant difference between AFC and age at culling in all groups  Breed and BW are influencers of timing of puberty  Conception rates are higher at third season from onset of puberty, so leave 2 months before timed artificial insemination from onset of puberty  It is more beneficial to reach target growth rates preweaning as this influences time to onset of puberty beyond ADWG after weaning  AFC of <23 months and >30 months results in lower MY  AFC of > 25 months gives a greater calving interval (no data less than 21 months)  There is no significant difference in first service conception rate from 23-30 months but overall fertility has been reported to decline from >26 months  As AFC increases above 22 months, survival tends to decrease but this is associated with production rather than the AFC

Limitations:
There was little mention of genetic influence on BW at birth and its consequent influence on first lactation yield but sufficient explanation was given regarding the small influence on heritability on AFC.

Population: Primiparous Holstein dairy cows
Sample size: Number of animals contributing to the dataset are not published. Data comes from 10 "large farms", which are described as >7,000 cows, from three regions in Iran but it is not clear exactly how many.
Intervention details: Using dairy software data to determine economic weight (in value)  ADWG was 0.2 kg/day greater preweaning on the intensive diet (0.64 vs 0.44 kg/day)  ADWG post-weaning was not significantly different  Age at conception and AFC was 15 days sooner in the intensive diet group  Milk yield (to 150 DIM and predicted 305 day yield) was not significantly different from each group  BCS at mating was not significantly different  Economic return was marginally in favour of intensive feeding but return could be positive or negative based on milk prices Limitations: Out of the n=40 heifers in each group, 38 (out of 80 total) were excluded (not known exactly how many from each group were excluded) for reasons of disease, late attainment of pregnancy, failure to conceive after three services, abortion, low MY and lameness. AFC was found to have no effect on yield but there were only 15 days between each group so differences were not significant.

Haworth et al. (2008) Australia
Population: Primiparous dairy cows Limitations: Very small sample number to make conclusions over effects of AFC on lifetime parameters: this is why yield in the first lactation had more effect on outcomes than AFC. One farm had huge variation in AFC: this implies that there may have been many other factors in heifer management that could have influenced the yield result later in life.  First lactation MY in lower quartile herds was 8,450 kg (SD 1,448) and 10,821 kg (SD 1,946) in higher quartile herds  Genetic associations of AFC and MY were negative regardless of herd "level", suggesting a genetic by environment interaction but the author states that this should be interpreted with caution as genetic relationships are dynamic within herds, especially in situations of population selection.  Lower quartile environment herds had genetic correlation of -0.31 and higher herds had one of -0.51 for both AFC and MY Limitations: Limited to herds contributing to the DHI US dataset and therefore may be better managed herds. Historical dataset (1987)(1988)(1989)(1990)(1991)(1992)(1993)(1994) and over multiple years, where many other factors may have influenced the dataset (especially in improvement programs). Linking genetic heritability to environmental factors and using AFC as an outcome measurement is very risky: AFC is influenced by every management factor in youngstock rearing, so linking genetics to this is extremely challenging. This is recognised by the authors as a limitation also.

Banos et al. (2007) United Kingdom
Population: Dairy cows  AFC was optimal between 22-24 months for first lactation yield, peaking at 7,000 kg (ranging 5,700 kg at 21 months to 6,100 kg at 34 months). This is true also for fat yield (190 kg for 22-24 months versus 170 kg for <22 months and 175 kg at 34 months)  AFC <22 months was detrimental for first lactation yield by up to 800 kg  AFC was not seen to directly influence longevity but productivity over lifetime positively influenced longevity.  Heritability of AFC was very low (0.082) Limitations: Data collected over a long period of time and many other factors could influence outcomes, especially if the herds are in improvement programs. Data from software recorded herds may be from better managed herds and therefore not be representative of the country population.   AFC ranged from 660-1156 days with peaks at 24 and 36 months  AFC decreased linearly, associated with first lactation yield of milk, fat and protein by 55.5 kg, 0.6 kg and 2.3 kg respectively for each month from 20 months  AFC was associated with survival to second, third, fourth and fifth parity  Animals with an AFC of 24 months had greatest odds of survival in the first lactation (odds ratio of around 1.0), decreasing to an odds ratio of around 0.7 for AFCs from 28-38 months  There was no association found with AFC and survival above the first parity Limitations: Collection of data over 6 years and other factors could influence the outcomes studied such as herd improvement programs and environment.
Herds specified as seasonal grazing herds but not specified whether spring or autumn calving. This may affect yield from grazing herds in Ireland as well as heat effects in Spring calving herds (even in Ireland it can reach over 24 °C). Very large range of AFC. Numbers of animals in distribution graphs peaked at 24 and 36 months (very far apart): this indicates a great variety of management factors that are influencing heifer growth, which may also affect the outcomes measured.

Changee et al. (2013) Korea
Population: Holstein dairy cows Lifetime profit had a very large residual spread over 3000 days (6 lactations) which could skew some of the conclusions on profitability for later lactations.  , 2004). Although this technique gives large sample population numbers, it may also create confounding factors: farms that use computer records and that participate in herd improvement schemes may not be representative of all dairy farms and are already skewing data towards the more advanced farm management systems. Therefore the AFCs recorded as the range within the country may be lower than the real picture. Also, if data from farms on improvement programs cover several years, it may be likely that the AFC and yields (and diets, disease management, heat abatement, etc.) may improve anyway, also confounding results.

Outcomes of AFC
Generally, MY data is well recorded on farms, as it represents the main form of income to the businesses. Only one of the reviewed papers mention the confounding effect of BW at calving on first lactation 305 day yield (Ettema & Santos, 2004). As AFC also influences first lactation 305 day yield, further research including this parameter would be preferable in future to attempt to separate these effects.
In contrast, fertility data is not so clear. Some papers use calving interval (the mean average between calvings in days); some use days open (the average number of days from calving to confirmed pregnancy) and others use first service conception rate (percentage of cows pregnant to the first service after voluntary waiting period). Unfortunately, calving interval is a very historical parameter: it generally represents cows that got pregnant at least 9 months to 1 year previous to the date it is recorded. Therefore, when measuring AFC effects in calving interval, the records may not be representative of the AFC group from the year studied, or include the confounding effect of culling in calculating calving intervals. Similarly, days open requires at least 85-120 days from calving to calculate a figure, so to be representative of the sample population, it must be recorded over at least a year. Again, this means that the AFC recorded may not be the same group of those providing the days open result. Also, when assessing improvements in days open, one must consider that it will go up before it goes down: only non-pregnant cows are included in the calculation: when a cow gets pregnant, she is removed from the "pool" of animals contributing to the data. Therefore, if more cows get pregnant, the pool of non-pregnant animals is skewed to a larger figure as it is more influenced by chronically non-pregnant cows. When these leave the population, the days open figure will finally go down but this may be a long time from the AFC figure recorded from the farm.
So the more representative figure for testing AFC is the first service conception rate used in Ettema  As far as profitability is concerned, in order to calculate the effect of AFC on yield and fertility, a considerable amount of modelling is required. There is a lot of potential here for different researchers to include different parameters into their model: feed prices change by country and over time, as does economy in general, therefore the conclusion of one paper may not be comparable with another. As the effects of AFC (yield, fertility and culling) are influenced by so many other factors, the model is required to have many inputs, from labour, feed price, veterinary costs, rent, energy costs, disease rates, etc. When this is calculated, it is easy for some groups to miss some inputs and also, without considerable detail, avoid double-counting. For this reason, it is not advisable to take a value published from these papers and directly translate that to the reader's situation.

Application and reflection
The clinical bottom line has far-reaching consequences for advisors on dairy farms. AFC is a key performance indicator of heifer management, including diet quality and availability, disease risk, insemination techniques and preweaning growth. A knowledge of the evidence that sets a target window of 22-26 months is a crucial tool in youngstock management advice. Not reaching these goals can be detrimental to both fertility and milk income, so hitting these targets gives both foundation and drive to improvement projects throughout the industry.
Further research in the interaction of body weight at birth and AFC on first lactation yield would provide more clarity on the individual effects of both on production.

Methodology Section
Three databases were used to search this PICO. The author has chosen to exclude papers before 2000, as dairy records before this date were limited and the management/genetics of dairy cows since this date have changed considerably in many countries through improvement schemes. One review Wathes et al. (2014) mentions papers back to 1979 and is a comprehensive work summarising much of the intervening period, so as it is included in this review, it was felt by the author to be a sufficient summary of the evidence pre-2000.
The PICO focused on effects rather than inputs to AFC so articles were excluded that concentrated on influences on AFC rather than consequences of varying AFC.
There have been publications linking AFC to genetic heritability traits, as this is an attractive theme to offer from genetic improvement (semen) companies. However, it has been evident from the literature that heritability of AFC is very low and not a significant factor (Ruiz-Sanchez et al. 2007): many other publications also make this conclusion. For this reason, the author has excluded genetic heritability papers from the Knowledge Summary.

Search Strategy
Databases searched and dates covered: