Impact of mineral feeding on reproductive efficiency on smallholder dairy farms in Kenya

Shauna Richards, John Vanleeuwen, Shepelo Getrude Peter¹, Jeff Wichtel², Collins Kamunde³, Fabienne Uehlinger⁴ and George Gitau¹

Centre for Veterinary Epidemiologic Research, Department of Health Management, Atlantic Veterinary College, 550 University Avenue, Charlottetown, PE, Canada, C1A 4P3
srichards@upei.ca
¹ Department of Clinical Studies, Faculty of Veterinary Medicine, University of Nairobi, Kenya
² Department of Population Medicine, Ontario Veterinary College, 50 Stone Road E., Guelph, ON, Canada N1G 2W1
³ Department of Biomedical Sciences, Atlantic Veterinary College, 550 University Avenue, Charlottetown, PE, Canada, C1A 4P3
⁴ Department of Large Animal Clinical Medicine, University of Saskatchewan, 52 Campus Drive, Saskatoon, SK, Canada, S7N 5B4

This project was supported by the Atlantic Veterinary College, Farmers Helping Farmers, the National Sciences and Engineering Research Council of Canada,
Universities Canada, World Agroforestry Agency, Veterinarians without Borders-Canada, and the University of Nairobi.

Abstract

A randomized controlled trial (RCT) on smallholder dairy farms in Kenya was conducted to evaluate the effect of enhanced post-partum feeding practices on the calving interval (CI) of cows. With initial biweekly RCT exams, the incidence rate of subclinical mastitis (SCM) was also determined. Privately owned cows (n=111) which had recently calved were randomly assigned to one of nine recommendation groups based on 3 feeding levels of commercial dairy meal and mineral (3x3 factorial structure): (1) 100% of daily recommended intake; (2) 50% of daily recommended intake; and (3) feeding on the farmer's discretion. Cows were evaluated biweekly for two months post-partum for SCM using the California Mastitis Test, and again after one and two years for a subsequent calving date. The incidence rate of SCM was 0.30 cases /cow-month. A total of 73% (81/111) of study cows calved or were confirmed >3 months pregnant over the two-year observation period, between 2013 and 2015. For those 81 cows, the mean and median CIs were 570 days and 519 days, respectively (range 306-913 days). A Cox proportional hazards model was used to evaluate the effect of feeding practices and other predictors on calving interval. Mineral feeding was very close to significant (p=0.052) when included as a time varying effect, with mineral feeding closer to the daily recommended intake leading to a higher hazard of calving, although the effect diminished as time went on. Taller, multi-parous, and high-producing cows had a significantly higher hazard of calving compared to shorter, primi-parous and low-producing cows, respectively. Farmers with college or university level education, and female farmers, had cows with a higher hazard of calving compared to farmers who did not finish primary education and male farmers, respectively. The final CI model did not include post-partum SCM.

Keywords: dairy cattle, mastitis incidence rate, nutritional management

Introduction

Dairy farming is an important source of income for many families living in high potential regions of rural Kenya (Wambugu et al 2011, Vanleeuwen et al 2012). Daily milk production (DMP) is typically low on smallholder dairy farms (SDF) in developing countries such as Kenya, and production limiting diseases can severely affect farm productivity and therefore the livelihoods of the farmers (Vanleeuwen et al 2012, Richards et al 2016). Production-limiting factors on SDF include mastitis and poor reproductive efficiency (Lyimo et al 2004, Vanleeuwen et al 2012). A better understanding of the frequency, distribution and factors associated with mastitis incidence and prevalence, and poor reproductive efficiency, would provide SDF with better evidence-based strategies for their managers and advisors to improve udder health and reproduction.

A majority of mastitis research has involved estimation of prevalence of infection through indirect (California Mastitis Test – CMT) or direct tests of Somatic Cell Counts (SCC) and/or milk bacterial culture results (Almaw et al 2008, Getahun et al 2008; Mdegela et al 2009, Abera et al 2012, Tolosa et al 2013, Gitau et al 2013). In Kenya, the prevalence of subclinical mastitis (SCM) identified with the CMT and confirmed with Petrifilms^TM on SDF in Mukurwe-ini and Nakuru districts was high, with 48-52% of cows being affected by prevalent cases (Gitau et al 2013). The most common isolate in that study wasStaphylococcus aureus (68% of cases), followed byStreptococcus agalactiae, and other Streptococci, Corynebacterium bovis, and Klebsiella spp. (Gitau et al 2013). Testing for SCC in Kiambu District found that 80% of cows had a SCC higher than 250 × 10³ cells/mL, and 43% had a SCC higher than 600 × 10³ cells/mL, with 63% of all cow samples culturing positive for pathogenic bacteria (Omore et al 1999). In other studies conducted in Ethiopia and Tanzania, the prevalence of subclinical and clinical mastitis ranged from 22.3%-62.0% and 1.80%-10.3%, respectively (Almaw et al 2008, Getahun et al 2008, Mdegela et al 2009, Abera et al 2012, Tolosa et al., 2013).

The cost of mastitis on SDF is substantial considering the high prevalence, production loss due to milk lost through the effects of infection, and milk lost from infected quarters due to antibiotic withdrawal times or undesired milk quality. In Kenya, Omore et al (1999) found that, on average, the daily loss in milk yield from a subclinical case was 0.20 kg/day, or a loss of 80.0 kg/cow/year, costing approximately $10.0 USD/cow/year (based on 1996 dollars - $25.0/USD/cow/year in 2016). A recent study in Mukurwe-ini, Kenya, found that cows in early lactation that were affected by SCM, based on CMT results, were associated with a loss of 0.3kg/day in milk production, among cows with average milk production of 13.0 kg/day (Richards et al 2016).

Research projects to determine incidence rates of subclinical and/or clinical mastitis on SDF are sparse because they require longitudinal monitoring programs of cow udder health. In East African reports, the incidence rate of clinical mastitis was found to be 31.7 cases/100 cow-years in Tanga and Iringa regions of Tanzania (Karimuribo et al 2006) and 43.3 cases/100 cow-years in Dar es Salaam, Tanzania (Kivaria et al 2007). Farmer-reported incidence studies are an alternative to longitudinal monitoring programs, but rely on accurate farmer recall and/or records. In Mukurwe-ini, Kenya, the incidence rate of farmer-reported mastitis was 0.550 cases/cow-year; however, this rate is reflective of clinical mastitis and subclinical cases (Vanleeuwen et al 2012). This rate could also be an over-estimate because it included cows assumed to have mastitis due to rejected milk, but without CMT confirmation.

Another cause of production losses in SDF is reproductive inefficiency. Reproductive inefficiency delays conception and subsequent calving, and reduces the amount of time spent at peak lactation, which leads to lower levels of milk production over the lifetime of a cow. It is generally considered that a 365 day calving interval provides the most efficient milk production (Esslemont 1993). In Kiambu District, Kenya, the average calving interval was 646 days; however, a shorter interval of 481 days was reported for farmers who were members of a dairy cooperative (Odima et al 1994). In Tanzania, long calving intervals have been reported at 500 days in the Usambara Mountains (Swai et al 2005), and 476 days in Coastal Tanzania (Swai et al 2007). Long calving intervals are related to long calving-to-conception intervals, which have been reported at 130 days in Tanzania (Lyimo et al 2004).

There is limited information on the factors associated with poor reproductive efficiency on SDF. Suboptimal nutrition will likely lengthen calving intervals on SDF, as cows fed a high energy diet postpartum conceived more quickly than cows fed a lower energy diet (Dunn et al 1969). Mineral and protein supplementation work synergistically to promote reproduction, maintenance, growth, and health of dairy cows (Abate 1998). In Africa, deficiencies in selenium and manganese are a general concern (Schillhorn et al 1990). Deficiencies in minerals in Kenyan cows may include: magnesium, phosphorus, sodium, copper, cobalt, and the appropriate ratio of calcium-to-phosphorus (Abate, 1998). Further study is needed to identify and quantify the main farm and cow factors associated with poor reproductive efficiency in dairy cattle on SDF.

The objectives of this study were, therefore, to determine the prevalence and incidence rate of subclinical mastitis on SDF during the first two months post-partum, to assess reproductive efficiency on SDF, and explore factors that influence reproductive efficiency.

Materials and methods

Study site and study population

This study was approved by the Research Ethics Board and the Animal Care Committee of the University of Prince Edward Island, the Mukurwe-ini Wakulima Dairy Ltd. (MWDL), and a partner non-governmental organization called Farmers Helping Farmers. Signed consent to join the study was obtained from all participants after the project had been fully explained.

The study was carried out in Mukurwe-ini, a sub-county of Nyeri County, Kenya, with members of the MWDL from June–September 2013 (early cool dry season to late cool dry season), with follow up in July 2014, and July 2015. There were over 6,000 members of the MWDL in 2013. Mukurwe-ini has an estimated population of 83,932 people as of 2009 and covers 179 km2 (Kenya National Bureau of Statistics 2010). Nyeri County is part of Kenya’s Central Highlands, spanning an area of 3,337 km2 (Kenya National Bureau of Statistics 2010). Mount Kenya is located to the east of Nyeri County at an altitude of 5199 meters, and the Aberdare Range is to the west, peaking at 3,999 meters (National Coordination Agency 2005). The study area is considered part of the humid highlands at an altitude of over 1,500 m, with annual rainfall over 1,000 mm and humidity >50% (Orodho 2006). This area is classified as an agro-climatic zone (I–III) that has a high potential for growing crops (Orodho 2006).

Study Design

Cows enrolled in this study were part of a two-month nutrition intervention trial that began in June, 2013 (Richards et al 2015, Richards et al 2016). Briefly, based on a sample size calculation to evaluate effect of feeding on milk production (Richards et al 2016), a goal of at least 108 newly calved cows (calved within 7 days of first visit to farm) were enrolled in a feeding trial and blocked by body condition score (BCS) into two groups; cows with a BCS equal or greater than 2.5/5, and cows with a BCS less than 2.5/5. After blocking, cows were randomly assigned to a factorial design incorporating three mineral feeding groups, and three commercial dairy meal (CDM) feeding groups, resulting in 9 feeding groups with 12 cows/group participating in that study. Random allocation to feeding groups was done by drawing from a hat after blocking based on BCS. Blinding of farmers to their treatment was not possible as they were required to implement the trial protocol. To ensure an adequate sample size, 111 farms were enrolled to allow for potential withdrawals from the study due to sale or death. Farms were contacted and enrolled from a database held by MWDL, consisting of cows that had been artificially inseminated (AI) by MWDL veterinary services between 9 and 10 months prior to the start date of the study. Farms were excluded if they had more than four adult cows as these farms were not considered representative of a typical SDF. Due to small herd sizes, only one cow per farm was enrolled. Farms all employed zero-grazing management, with some form of stall/bedding area, and a manger/eating area.

The CDM was formulated for lactating cows and produced by Bora Feeds, Ichamara, Nyeri County (Table 1). Based on a formula (kg CDM = (kg DMP – 5)/ 2) utilizing daily milk production levels (Richards et al 2016), the three levels of CDM feeding were: (1) full recommended daily intake; (2) 1/2 recommended daily intake; and (3) no recommendation made (intake at the discretion of the farmer). The mineral feeding was similarly assigned based on milk production: (1) full recommended daily intake; (2) 1/2 recommended daily intake; and (3) no recommendation made (intake at the discretion of the farmer). Daily recommended mineral intake was based on the feeding directions of the manufacturer of the mineral provided to cows (Maclik Super, Cooper K Brand Ltd) (Table 2).

A second sub-population of 57 lactating cows that were herd-mates of the 111 study cows were also assessed for SCM during the study period. These herd-mates did not have information collected beyond CMT and milk cultures. Their inclusion in this study was only to describe SCM incidence and prevalence and culture results among cows on the same study farms but at a different stage of lactation, for comparison purposes.

During the study period, cows were visited every two weeks for a total of five visits. On each visit, a CMT test was performed and scored as 0, trace (T), 1, 2, or 3 on each milking quarter. Cows with CMT scores ≥1 were considered to have subclinical mastitis. Cows with visibly abnormal milk, positive CMT, and other clinical signs, such as swollen or painful udder, and/or endotoxin-induced shock consistent with mastitis, were diagnosed as having clinical mastitis. An incident case of subclinical mastitis was defined as a CMT ≥1 if the past CMT result was a zero, or a CMT result ≥2 if the past CMT scored as trace. A quarter classified as mastitic was considered cured when the CMT result of that quarter returned to zero or trace (T) following treatment. The cured quarters were then considered “at-risk” for new cases of mastitis.

Cows that had mastitis on the first visit, or developed a new case of mastitis on subsequent visits, had a milk sample taken for bacterial culture. When a new incident case was noted in a quarter(s) where a cow also had a prevalent case of mastitis in another quarter(s), then, a second sample was not taken for budgetary reasons, assuming the pathogen of the new infection was likely the same as the pathogen of the prevalent infection. If a new case of mastitis occurred following cure of all quarters, then a new sample was taken. Milk was pooled from affected quarters if more than one quarter was affected, provided the newly infected quarters were identified at the same visit. Milk samples were kept on ice packs until they could be frozen at -18⁰C; once frozen, they were transported to the University of Nairobi Veterinary College for routine mastitis culture and sensitivity (Gitau et al 2013, Gitau et al 2014).

Aside from culture and CMT, other data collected on the study cows on the first visit included: farmer demographics (e.g. age, gender, education level), and cow history of disease during the last lactation, height (at the withers) and signalment parameters (e.g. breed and parity). At each of the five visits, data collection also included: farm- and cow-level feeding practices (Richards et al 2016), and cow parameters, such as weight (via weight tape), physical exam findings, and BCS (Rodenburg 2012).

To assess reproductive efficiency in these cattle, the cows were re-visited in July of 2014 and 2015. During these visits, cows were examined for their reproductive status but not for SCM since incidence could no longer be determined and prevalence was no longer of interest. Feeding practices information and cow parameters were collected, including results of any pregnancy test performed by a veterinarian prior to the visit, and/or a record of calving or abortion since the last visit. This allowed for determination of calving interval (time between two subsequent calvings, the first being in 2013).

Descriptive Statistical Analyses

Descriptive statistics were calculated for the data collected on the initial visit in 2013, at each monitoring visit during 2013, and/or subsequent visits in 2014 and 2015, as applicable. The relevant data were assessed for the following outcomes of interest: the cultured organisms causing SCM, SCM prevalence risk, SCM incidence rate, and calving interval.

For SCM incidence, prevalence, and culture results, the two groups of cows (newly calved cows, and all other cows) were assessed separately. For each group, the data from each visit were collapsed to give a cumulative count of the number of incident cases of mastitis/cow at the quarter level for that visit. With the biweekly sampling in 2013, a quarter was considered no longer at-risk for incidence rate calculations during the week prior to a new case being identified on CMT (i.e. a case of mastitis lasted for a duration of one week, on average, and therefore a new case could not occur in that quarter during that time period). For cured quarters, the cured quarter was considered at-risk for incidence rate calculations for the week prior to the quarter being identified as cured. The incidence rate was calculated at the quarter level per cow-day and then converted to the cow level using cases per cow-day and cow-month units.

The calving interval for each cow was calculated on the basis of the presence of a live calf or reporting of a live calf as indicated by the farmer, or by pregnancy diagnosis of a pregnancy ≥ 3 months duration by a veterinarian (and in some cases, projecting a calving date if the cow was diagnosed ≥ 3 months pregnant at the 2015 visit).

The cow data from each 2013 visit were collapsed to give a cumulative sum of the number of times farmers indicated they fed high protein forages on a given visit, and an illness score (a response of yes (1) or no (0) to a set of physical exam parameters, including retained placenta and metritis). The mean of the daily milk production, CDM fed/day, and mineral fed/day were also calculated.

Feeding data were analyzed based on how farmers actually fed their cows as opposed to the assigned feeding amounts or groups because these were privately owned cattle with owners who would alter what they fed to their cows because of misunderstandings in recommendations or lack of compliance. Other predictor variables examined for calving interval analytical statistics included the following: weight, height, parity, and body condition score at the first visit; breed of the cow; history of having mastitis in the last lactation; history of illness in the last lactation or dry period leading up to this lactation; if the cow received CDM or mineral in the month prior to calving; the size of the herd; size of the household of the farm; proportion of income from dairy farming; gender of the principal farmer; education level of the principal farmer; age of the principal farmer; and size of the farm.

Modelling of Calving Interval

The outcome of interest utilized to determine factors of reproductive efficiency was the calving interval (i.e. the period in days between the 2013 calving date and the subsequent calving date) for the 111 nutrition trial cows. To model these data, a Cox proportional hazards model was utilized (survival analysis). The 57 herd-mates were not included in this analysis because they were not part of the nutrition trial, and no follow-up reproductive data were collected for these cows. Predictors which were considered as known or possible confounding variables were considered for inclusion in the model. The following transformations of the CDM and mineral feeding variables were: 1) on each visit, the daily intake of CDM and mineral were calculated for each cow as proportions of the cows daily requirement (0-100% daily requirement) based on their milk production; and 2) an average proportion over the 5 measurement periods was calculated for each of the transformed nutrition variables.

Cox proportional hazards models were initially used to screen univariable associations with p values of <0.25, and then a final Cox model was built with any confounders and variables with a final significance of p-value of < 0.05. The model was assessed for proportional hazards, assumption of independent censoring, overall fit of the model, functional form of the predictors, and assessment for outliers and influential points. In all models, the Efron method was used to handle ties.

Results

Descriptive Statistics

All descriptive statistics of variables considered for inclusion in modelling are found in Richards et al (2015) and Richards et al (2016). Briefly, all farms were zero-grazed and small in land area; the median farm was 1.50 acres, and the largest farm was 10.0 acres. The primary farmer was usually married (85.0%), and half of the farms were run by women. Financially, 62.0% of farmers got at least half of their yearly income from dairy farming. Cows were mostly Holstein crossbreds (90/111), and the rest were other dairy crossbreds (21/111). The average weight of cows was 378 kg, and average height was 122 cm.

Of the 111 nutrition trial cows, 41 cows had no incident cases of SCM, and only 26 cows had no prevalent cases of SCM (Table 3) during the 8-week monitoring period post-calving. Three cows had 10 prevalent and incident cases of SCM during this time frame, half of which were incident cases and half of which were carry-overs from the previous 2 week sampling. The raw incidence rate of SCM was 0.010 cases/cow-day (95% CI 0.009-0.014) or 0.300 cases/cow-month.

The 57 herd-mates of the 111 nutrition trial cows had an incidence rate of SCM of 0.008 cases/cow-day (95% CI 0.006 – 0.013), or 0.240 cases/cow-month. The mean days-in-milk for this group of 57 cows at the start of the monitoring period was 292 days (range 16 -1121 days).

A total of 96 samples were cultured from the 139 incident cases from the 111 nutrition trial cows and the 57 herd-mates. A total of 102 samples were taken from cows on the farms, but 6 samples were not cultured in the laboratory for logistical reasons.

The most common culture result was Staphylococcus aureus, followed by no growth, and Streptococcus agalactiae (Table 4). There were some differences in pathogens between the 111 recently calved cows and their 57 herd-mates in later lactation; the herd-mates had the only cases of E. coli, S. dysgalactiae, and mixed growth. However, the herd-mates had proportionately more cases of S. aureus compared to the 111 early lactation cows; however this was not statistically significant (p > 0.05).

Seventy-three percent (81/111) of study cows calved or were confirmed >3 months pregnant over the two-year observation period, between 2013 and 2015. For those 81 cows, the mean and median calving intervals were 570 days and 519 days, respectively, with a range of 306 to 913 days. The remaining 30 cows that did not calve were monitored for an average of 452 days (range 60-767 days), with cows in the low and medium parts of this range being lost to follow-up between the annual assessments. The median calving interval for multiparous cows and first-calf heifers was 506 days and 608 days, respectively (Table 5). Median CIs by gender and education level of the primary farmer are also found in Table 5.

Calving Interval Survival Analysis

The following variables were found to have univariable associations with calving interval (p < 0.25): cow height, breed, and parity; initial BCS; the average daily milk production, cumulative number of reports of feeding of high protein forages, comfort score of stall during the 2 month monitoring period in 2013; history of abortion; and the gender, education level, and age of the primary farmer.

In the final Cox survival model, four variables were found to be significantly associated (p < 0.05) with calving interval, and two variables were very close to significant (p = 0.052, 0.052) (Table 5 and 6). One cow had no recorded height, and since height was a significant variable in the final Cox model, there were only 110 cows in the final model. The descriptive statistics in Table 5 show that the final model was based on mostly multiparous cows of a short stature and modest milk production, managed by farmers of equal gender but education levels limited to high school or less. Average daily milk production among the 111 cows during the first 2 months post-partum was only 13.4 kg/cow/day, ranging from 6.50 to 25.6 kg/cow/day, with average proportion of daily required intake of mineral supplement being lower than recommended for most cows (Table 5).

Interpreting the final model (Table 6), multi-parous cows had a 1.99 times higher hazard of calving than first calf heifers (parity 1) that was very close to significant (p=0.051). There was a significant curvilinear effect of height on the hazard of calving, with taller cows having a higher hazard of calving than shorter cows; however, this association plateaued at the upper range of heights. As average daily milk production increased during the 2 month monitoring period, the hazard of calving increased by 1.09 for every additional kg of milk per cow per day. Farmers with a college or university level of education had a 17.3 times greater hazard of their cow calving then those not finishing primary education; however, there was high variability of the hazard ratio for those with the highest level of education, and a small sample size in this group (n=5). Female farmers had a 1.95 higher hazard of their cow calving then male farmers. The proportion of daily required mineral fed per day was very close to significant when included in the model as a main effect and a time varying effect. The overall effect of this variable was that higher mineral feeding increased the hazard of calving; however, this association decreased over time. Goodness-of-fit tests (e.g. proportional hazards, independent censoring, assessment for outliers and influential points) demonstrated a good model fit.

Discussion

This study provides a better understanding of the descriptive statistics and factors associated with subclinical calving interval on SDF in Kenya, with better evidence-based strategies for their managers and advisors to improve reproduction. These results were part of secondary objectives of a nutrition trial determining the milk production effects of better dairy meal and mineral feeding on post-partum cows, since it was theorized that cows fed better during the trial would likely also have better reproduction. Also, since we were on the farms every two weeks during the post-partum period to determine milk production and feeding management, we were able to conduct a CMT every two weeks to obtain data on incidence of SCM, which is more useful information for understanding disease frequency and transmission than prevalence data.

Early lactation cows on SDF in Mukurwe-ini, Kenya, had an incidence rate of SCM of 0.30 cases/cow-month, and a mean calving interval of 571 days in the 73.0% of cows which had a subsequent calf by the second annual visit. These findings are indicative of the presence of the production limiting factors facing dairy farmers in Mukurwe-ini, Kenya.

The most common culture result for the incident and prevalent cases of SCM in our study was Staphylococcus aureus, followed by no growth, and Streptococcus agalactiae (Table 4). For prevalent cases of SCM among cows on SDF in Nyeri and Nakuru Districts in Kenya, similar contagious pathogens were reported (Gitau et al 2013), with limited numbers of SCM culturing coliform pathogens. In both studies, samples were frozen prior to culturing for logistic reasons (farms were far from the laboratory). For both contagious and environmental pathogens, total bacterial counts are decreased after freezing, however coliform pathogens are especially negatively affected by freezing (Alrabadi 2015), which could explain why few samples cultured positive for coliform pathogens. It is probable that a portion of the no growth samples were coliform infection that had either self-cured already, or did not grow due to freezing effects

With the long CI, we were able to detect important farmer-demographic, cow-level, and farm-management factors associated with the hazard of calving in the SDF context, even with the relatively small sample size (Table 6). Cows with more than one calving at the study onset had a trend of a higher hazard of calving (1.99) in the two-year study period. This predictor was very close to statistically significant, and therefore was included in the final model but should be interpreted with caution. First parity cows are usually still growing, leading to a greater potential for negative energy balance, impeding reproductive efficiency.

Taller cows also had a higher hazard of calving, although this association plateaued at the tallest heights reported (Table 6). There is no biologic reason that a taller cow should conceive faster, however taller cows may be better quality purebred dairy animals that were raised well, and are also fed well as lactating cows, making them more likely to have a better foundation to conceive and calve earlier. Additionally, farmers who made an investment in improved animals could be motivated to get these animals pregnant in order to make more money.

Higher average daily milk production in early lactation was associated with an increased hazard of calving (1.09), which may seem counter-intuitive because higher milk production could lead to greater negative energy balance, impeding reproductive efficiency (Table 6). However, milk production was used to determine the amount of CDM to feed, therefore high-producing cows were frequently receiving more CDM, perhaps explaining this finding. With milk production being correlated with CDM feeding recommendations, and milk production being significantly associated with calving interval in the model, it is not surprising that CDM feeding was not found to be significantly associated with the hazard of calving. The small sample size could also have limited the number of variables in the final model.

While some feeding practices, such as concentrate feeding and use of high protein forages, were not significantly associated with the hazard of calving in the final model, feeding mineral was very close (p = 0.052) to being significantly associated with the hazard of calving and therefore was included in the final model (Table 6). When cows which were fed more mineral to ensure daily required needs were more closely achieved, these cows had a higher hazard of calving than those fed lesser amounts of mineral. This effect changed over time; as time went on, the effect of mineral feeding was reduced. This is not unexpected, as the effect of mineral feeding is expected to have a positive effect on cows who are otherwise healthy and able to conceive, but are mineral deficient (Abate 1998). For cows that are unhealthy or unable to conceive for other reasons besides mineral deficiency, then they will seem to be refractory to feeding of mineral feeding on reproductive outcomes. Ensuring sufficient trace minerals to cattle has been shown to improve fertility in a clinical trial elsewhere (Mundell et al 2012). With the average proportion of daily required intake of mineral supplement being 56.7% (Table 5), there is ample room for improvement in mineral feeding among the study farms. Some of the farmers (>1%) in the mineral group designated “no recommendation made (intake at the discretion of the farmer)” gave their post-partum cow no supplemental mineral, shown by the low end of the range being zero (Table 5), which could be a function of limited knowledge on mineral feeding or limited funds to purchase mineral.

Farmer-related factors associated with the hazard of calving were education levels and gender (Table 6). Farmers with college or university education had a significantly higher hazard (17.3) of having their cows calve than farmers with lower levels of education (there was no significant difference in the hazard of cows calving between the two lower levels of education of farmers). Caution should be used in interpreting this effect, as only 5 farmers had higher education, and the 95% confidence interval was wide (3.86-77.8), however, it would stand to reason that more educated farmers are more knowledgeable about the importance of shorter calving intervals and how to attain them. In Bangladesh, smallholder farmers with secondary or higher levels of education were 9.70 times more likely to adopt new farming technologies then illiterate farmers (Quddus 2012). The farmers in the study run in Bangladesh who adopted new technologies, such as AI, improved genetics, and farming practices, had higher daily milk production then those not practicing improved technologies. Quddus (2012) also demonstrated a correlation between household income and adoption of new technologies, indicating the socioeconomic status of farms can affect the productivity of SDF through adoption of new technologies.

While not noted in our study at the MWDL, Quddus (2012) also found that age of the farmer and the farm size were correlated with adoption of new technologies, with older farmers adopting fewer new technologies, and farmers with larger farms also adopting fewer new technologies. While these correlations were noted, Quddus indicated that farmers, regardless of their education and other demographics, are all faced with challenges such as: high price of inputs, no technical assistance or extension services, and lack of awareness of new technologies to implement on their farms.

Female farmers had a higher hazard (1.95) of their cows calving than male farmers, perhaps due to female farmers spending more time with their cows and being more perceptive in identifying heat in their cows than male farmers (Table 6). On peri-urban Kenyan dairy farms, it was noted that female farmers did more dairy farming activities then male dairy farmers, even though male farmers had more control over farm resources (Kimani et al 2007).

In this dataset, very few cows had retained placenta or metritis (n=5), making them poor predictors in a model. However, in larger datasets and in a clinical setting, these uterine diseases are factors which should be considered to reduce conception and extend CIs (Swai et al 2005). Conversely, many cows had mastitis in this study, which may be why no association was seen between illness score and the hazard of calving. The data collected on reproductive efficiency were secondary objectives to the nutrition trial, and therefore, the sample size may have been inadequate to detect other associations of farm management and cow factors on the hazard of calving.

While not evaluated in this study, membership to a dairy group with improved services was found to reduce calving intervals from 646 days to 481 days in Kiambu District Kenya (Odima et al 1994). Having access to AI, veterinary care and animal feeds on credit were suggested to be the reason that member farmers had shorter calving intervals than non-members. In our study in Mukurwe-ini, all farmers were members of the MWDL and had access to the aforementioned services; however, their mean calving interval of 571 days was longer than the previously mentioned study, for unknown reasons.

One limitation of our study was that accurate data collection on reproductive performance was difficult to obtain because farmers in this study failed to keep reproductive records, and many farmers had a hard time remembering when their cow first came into heat following calving. Therefore, we chose to focus on calving interval as a reasonable and reliable reproductive outcome. When farmers were questioned about pregnancy status of their cows one year after the nutrition study, many farmers indicated their cow was pregnant and had been confirmed pregnant. However, in the second year of follow-up, many of these cows had not calved or calved much later than originally anticipated. It was not uncommon that these same cows had been bred again since the first year follow-up visit, meaning the originally believed pregnancy confirmation was either false or early embryonic loss had occurred in the original pregnancy. Due to these issues, the research team made sure to diagnose the pregnancy themselves, or view a calf on the farm in the second year of follow-up to confirm calving interval data accuracy.

While calving interval as the outcome was accurate for most cows, in some instances, farmers’ cows may have started to cycle fairly early after calving but were still unable to conceive successfully or at all. For example, one cow came into heat one month after calving but because of an undiagnosed uterine infection, she only conceived after treatment for the uterine infection eight months later. A few farmers also reported that they chose not to breed their cow, either due to lack of funds, or because they planned to sell the cow after the current lactation. Additionally, farmers who have a difficult time detecting heat (e.g. due to lack of knowledge or resources) may have also contributed to longer calving intervals or not calving at all in the two year follow-up period.

In the future, research should be done to evaluate the effect of environmental hygiene and milking practices on the incidence rate of mastitis to reduce the high rates seen in this study. Additional work on the incidence rate of SCM of cows in various stages of lactation, and their causative factors, would also be helpful to determine if stage of lactation plays a role in the rate and causes of SCM incidence.

Conclusions

The incidence rate of subclinical mastitis is high in cows on smallholder dairy farms in Kenya, with ample opportunity to improve udder health through reduced transmission. Further research is required to evaluate factors affecting the incidence rate, so that appropriate recommendations can be made to farmers to reduce the negative impact of mastitis on farm productivity.

Long calving intervals were also noted in the study cows. Feeding mineral in closer approximation to the daily required intake reduced calving intervals, especially in the early post-calving period. Farmers with higher levels of education, and taller cows also had reduced calving intervals. Future work to reduce calving intervals should focus on farmer education on proper mineral feeding practices, post-calving management, and growth of youngstock replacements with good genetics.

Poor record-keeping was noted among the farmers. Farmers in Kenya and similar settings should be trained on the benefits of good record-keeping in order to monitor mastitis and reproductive events, among other topics. Farmers should be advised to record all heats, breeding dates, and confirmation of pregnancy by a veterinarian or animal health worker. With the limited education of the farming population in Kenya, continued training on best management practices related to feeding, heat detection and timing of AI would likely also be of benefit to reduce reproductive inefficiency. These types of practices, along with improved SCM prevention, should lead to improved productivity on farms, as well as higher quality milk to drink and sell to cooperatives and processing facilities.

References

Abate A 1988 Effect of pasture mineral levels on extensive cattle production in Kenya, In Pasture Network for Eastern and Southern Africa (PANESA). African forage plant genetic resources, evaluation of forage germplasm and extensive livestock production systems. Proceedings of the Third Workshop held at the International Conference Centre Arusha, Tanzania, April 27-30,1987. ILCA, Addis Ababa, retrieved May 14, 2018, from http://www.fao.org/wairdocs/ilri/x5491e/x5491e19.htm

Abera M, Habte T, Aragaw K. Asmare K and Sheferaw D 2012 Major causes of mastitis and associated risk factors in smallholder dairy farms in and around Hawassa, Southern Ethiopia. Tropical Animal Health and Production. 44, 1175-1179.

Almaw G, Zerihun A and Asfaw Y 2008 Bovine mastitis and its association with selected risk factors in smallholder dairy farms in and around Bahir Dar, Ethiopia. Tropical Animal Health and Production, 40, 427-432.

Alrabadi N I 2015 The effect of freezing on different bacterial counts in raw milk. International Journal of Biology, 7, 9-12.

Bundi R M 2011 The prevalence, risk factors, antibacterial sensitivity and usefulness of Petrifilms as a diagnostic tool for bovine mastitis in Mukurwe-ini and greater Nakuru Districts, Kenya. MSc. Thesis, University of Nairobi, Kenya.

Dunn T G, Ingalls J E, Zimmerman D R and Wiltbank J N 1969 Reproductive performance of 2-year old Hereford and Angus heifers as influenced by pre- and post-calving energy intake. Journal of the American Veterinary Medical Association, 29, 7l9-726.

Esslemont R J 1993 Relationship between herd calving to conception intervals and culling rate for failure to conceive. Veterinary Record, 133, 163-164.

Getahun K, Kelay B, Bekana M and Lobago F 2008 Bovine mastitis and antibiotic resistance patterns in Selalle smallholder dairy farms, central Ethiopia. Tropical Animal Health and Production, 40, 261-268.

Gitau G, Bundi R M, Vanleeuwen J and Mulei C M 2013 Evaluation of Petrifilms^TM as a diagnostic test to detect bovine mastitis organisms in Kenya. Tropical Animal Health and Production, 45, 883-886.

Gitau G, Bundi R M, Vanleeuwen J and Mulei C M 2014 Mastitogenic bacteria isolated from dairy cows in Kenya and their antimicrobial sensitivity.Journal of the South African Veterinary Association, 85, 1-8.

Karimuribo E D, Fitzpatrick J L, Bell C E, Swai E S, Kambarage D M, Ogden N H, Bryant M J and French N P 2006 Clinical and subclinical mastitis in smallholder dairy farms in Tanzania: Risk, intervention and knowledge transfer. Preventive Veterinary Medicine, 74, 84-98.

Kenya National Bureau of Statistics 2010 Kenya 2009 Population and Housing Census –Volume 1B –Population Distribution by Political Units, retrieved May 14, 2018, from https://www.knbs.or.ke/publications/

Kimani,V N, Ngonde A M, Kang’ethe E K and Kiragu M W 2007 Gender, perceptions and behaviour towards health risk associated with urban dairy farming in Dagoretti Division, Nairobi, Kenya. East African Medical Journal, 84, S57-64.

Kivaria F M, Noordhuizen J P T M and Msami H M 2007 Risk factors associated with the incidence rate of clinical mastitis in smallholder dairy cows in the Dar es Salaam region of Tanzania. Veterinary Journal, 173, 623-629.

Lyimo Z C, Nyka R, Schoonman L and van Eerdenburg F J C M 2004 Post-partum reproductive performance of crossbred dairy cattle on smallholder farms in sub-humid coastal Tanzania. Tropical Animal Health and Production, 36, 269-279.

Mdegela R H, Ryoba R, Karimuribo E D, Phiri E J, Løken T, Resken O, Mtengeti E and Urio N A 2009 Prevalence of clinical and subclinical mastitis and quality of milk on smallholder dairy farms in Tanzania. Journal of the South African Veterinary Medical Association, 80, 163-168.

Mundell L R, Jaeger J R, Waggoner J W, Stevenson J S, Grieger D M, Pacheco L A, Bolte JW, Aubel N A, Eckerle G J, Macek M J, Ensley S M, Havenga L J and Olsen K C 2012 Effects of prepartum and postpartum bolus injections on trace minerals on performance of beef cows and calves grazing native ranges. The Professional Animal Scientist, 28, 82-88.

National Coordination Agency (NCA) for Population and Development, Ministry of Planning and National Development 2005 Nyeri District Strategic Plan 2005-2010 for Implementation of the National Population Policy for Sustainable Development. Nairobi, Kenya, retrieved May 14, 2018, from http://zawadivillages.wikispaces.com/file/view/Nyeri+plan+for+devt.pdf.

Odima P A, McDermott J J and Mutiga E R 1994 Reproductive performance of dairy cows on smallholder dairy farms in Kiambu District, Kenya: Design, methodology and development consideration. Kenya Veterinarian, 18, 366-368.

Omore A O, McDermott J J, Arimi S M and Kyule M M 1999 Impact of mastitis control measures on milk production and mastitis indicators in smallholder dairy farms in Kiambu District, Kenya. Tropical Animal Health and Production, 31, 347-361.

Orodho A B 2006 Country Pasture/Forage Resource Profiles Kenya. Rome: FAO, retrieved May 14, 2018, from https://pdfs.semanticscholar.org/1d7b/3d82527bce586638645a32e68f9df501b759.pdf\

Quddus M A 2012 Adoption of dairy farming technologies by small farm holders: practices and constraints. Bangladesh Journal of Animal Science, 41, 124-135.

Renaud J K, Seykora A J, Heins B J, Endres M I, Farnsworth R J and Bey R F 2005 Association between hygiene scores and somatic cell scores in dairy cattle. Journal of the American Veterinary Medical Association, 227, 1297-1301.

Richards S, Vanleeuwen J, Shepelo G, Gitau G K, Collins C K, Uehlinger F and Wichtel J 2015 Associations of farm management practices with annual milk sales on smallholder dairy farms in Kenya. Veterinary World, 8, 88–96.

Richards S, Vanleeuwen J, Shepelo G, Gitau G K, Wichtel J, Kamunde C and Uehlinger, F 2016 Randomized controlled trial on impacts of dairy meal feeding interventions on early lactation milk production in smallholder farms of Central Kenya. Preventive Veterinary Medicine, 125, 46-53.

Rodenburg J 2012 Body Condition Scoring of Dairy Cattle. Ontario Ministry of Agriculture, Food and Rural Affairs, retrieved May 14, 2018, from http://www.omafra.gov.on.ca/english/livestock/dairy/facts/00-109.htm

Schillhorn van Veen T W and Loeffler I K 1990 Mineral deficiency in ruminants in SubSaharan Africa: A review. Tropical Animal Health and Production, 22,197-205.

Swai E S, Bryant M J, Karimuribo E D, French N P, Odgen N H, Fitzpatrick J L and Kambarage D M 2005 A cross-sectional study of reproductive performance of smallholder dairy cows in coastal Tanzania. Tropical Animal Health and Production, 37, 513-525.

Swai E S, Kyakaisho P and Ole-Kawanara M S 2007 Studies on the reproductive performance of crossbred dairy cows raised on smallholder farms in eastern Usambara mountains, Tanzania. Livestock Research for Rural Development, retrieved May 14, 2018 from http://www.lrrd.org/lrrd19/5/swai19061.htm

Tolosa T, Verbeke J, Piepers S, Supré K and De Vliegher S 2013 Risk factors associated with subclinical mastitis as detected by California Mastitis Test in smallholder dairy farms in Jimma, Ethiopia using multilevel modelling. Preventive Veterinary Medicine, 112, 68-75.

Van Leeuwen J A, Mellish T, Walton C, Kanaru A, Gitau G, Mellish K, Maina B and Wichtel J 2012 Management, productivity and livelihood effects on Kenyan SDF from interventions addressing animal health and nutrition and milk quality. Tropical Animal Health and Production, 44, 231-238.

Wambugu S, Kirimi L and Opiyo J 2011 Productivity Trends and Performance of Dairy Cows in Kenya, Tegemeo Institute of Agriculture Policy and Development, Nairobi, Kenya, retrieved May 14, 2018 from http://www.tegemeo.org/index.php/resources/data/167-policy-briefs/465-productivity-trends-and-performance-of-dairy-farming-in-kenya.html

Received 16 November 2018; Accepted 4 May 2019; Published 4 June 2019

Go to top