High Production and Health: A Curious Paradox

Posted on May 7th, 2007

Title	:	High Production and Health
Source	:	University of Minnesota
Author	:	John Fetrow, DVM
Date	:
Content	:

High Production and Health: A Curious Paradox

John Fetrow VMD, MBA
College of Veterinary Medicine, University of Minnesota, St. Paul, MN

Steve Eicker DVM, MS
Valley Agricultural Software, King Ferry, NY

The Paradox

There is a paradox that pervades dairy farming, a paradox not often spoken about in simple terms but one that shapes many decisions made by dairymen and those who advise them.

1. On one hand, it is taken as self-evident that if dairymen find ways to remove stressors from their cows, the cows will respond by being healthier and producing more milk. Dairymen work hard to produce better-balanced rations, keep their cows more comfortable, cool cows in hot weather, dip teats and vaccinate to prevent disease, improve ventilation, and provide clean accessible water. All these things are done with the goal of providing a stress free life for their cows. We all accept that if these things are done well, there will be less disease, fewer “broken” cows, and less need to replace cows (cull) that have been damaged to a point of low economic value. In short, it is generally accepted that cows less stressed are less at risk and higher producing cows.

2. On the other hand, there is also a strongly held belief that cows that are high producers are more stressed than their low producing herd mates. High producing cows are believed to be more at risk of clinical disease and other negative consequences of high production that will ultimately lead to enough damage and will cause the cow to “crash” in production or to be culled (or die). If asked, dairymen will offer the opinion that high production means that cows will breed back more poorly, have more mastitis, suffer more metabolic and other disease, and be more likely to be culled. High producing cows are described as being “stressed”, “walking a tightrope”, or “on the edge”. Dairymen express concerns about “burning out cows” and talk of not “pushing their cows” for higher production. The short version of this belief might be termed “high production breaks cows”.

The paradox is obvious: if cows with lower stress become higher producers, how can it be true that high production stresses and “breaks” cows?

Consider the first part of the paradox. There is an enormous array of science to support that as we better care for the cow (feed, environment, disease control, comfort, etc.), she responds by consuming more feed and making more milk. As we remove the “bottlenecks” (stressors) that inhibit her from expressing her true genetic potential, the cow is “freed” to improve milk production and suffers fewer disease events. It certainly makes some sense that higher producing cows ought to be the ones that are the most healthy and least stressed.

At the same time, it is equally obvious that sometimes cows are hurt by management actions intended to increase production. An obvious example might be over feeding grain with the goal of increasing milk production. In the short term this strategy might be successful, but in the long term the effects on foot and rumen health make it counterproductive. In this example, it is not the high production per se that led to the cows’ problems, but an error of management. Had the cows achieved the higher production by another route (increased access to water, for example), it might be that there would be no health effects on the cow or even improved health along with higher production.

Avoiding Logical Leaps over the Cliff of Causality

If there is a wealth of support for the first part of the paradox (reducing stress increases production), where is the source of the belief in the second, contradictory part? We will assert that the second part of the paradox derives largely from the general observation that as the dairy industry has changed over the past several decades, as production per cow has increased, we have also seen what are believed to be notable increases in certain types of problems in dairy cows. With the two observations (more milk, more problems) appearing concurrently, it is only natural to believe that the two are linked and in fact that the increased production is the cause of the increased level of certain problems. Nearly everyone falsely concludes that because we have seen an upward trend in certain clinical problems across the industry as production levels across the industry have increased, that it follows that an individual cow is more at risk as her individual production level increases. In general, there is little data to support that this is true, but it is nonetheless widely believed and is the prevailing assumption across the industry.

Consider an example of the changes in the dairy industry over the past several decades. Figure 1 shows Minnesota annual summary data for herds from 1950 to 2001: 50 years of history (MN DHIA). It shows a clear association between the increasing production of cows on DHIA test and the cull rate as measured by the proportion of the average herd size that exited the herd each year. Statistically speaking, the correlation between culling and production is incredibly tight (R2of 0.88, p<<<.01). What does the correlation mean? Does it “prove” that high producing cows are more likely to be culled? Does this make sense? Do most dairymen choose to cull more high producing cows than low producing cows? In fact, both common sense and published papers show that higher production “protects” cows from being culled (Groehn 1998). How about viewing the data from the other perspective? Does the high (and statistically significant) correlation “prove” that increasing culling increases production? Or as a final interpretation, does this graph only establish that increased production and increased culling happened at the same time, and doesn’t necessarily prove anything about any causal connection?

In epidemiology, falsely concluding that two events occurring simultaneously must be causally connected is referred to as an “ecological fallacy”. Miles traveled by air in the United States have gone up steadily over the past several decades, as have the number of white tailed deer killed on the highway. Few of us would consider these two concurrent trends to be causally connected. Similarly, after a little thought, most of us would remain skeptical of an interpretation of Figure 1 as “proving” that increased production in a cow makes her more likely to be culled. But consider Figure 2. This figure presents Minnesota DHIA data for the past two decades, showing milk per cow and average days open. Again, there is a tight correlationbetween the two parameters (statistically significant correlation, R2= 0.80, p<0.0002). What does this correlation mean? Actually, it means no more than the correlation of culling and milk production did (or air miles and dead deer). It only means that the two things happened at the same time. Unfortunately, nearly everyone in the dairy industry would leap to the conclusion that Figure 2 (and the accompanying statistical testing) “proves” that higher production “causes” poor reproduction. Couldn’t we just as easily conclude that poor reproduction “causes” higher production and start breeding cows when they are not in estrus to increase milk output? There might in fact be a causal connection between increased production and poorer reproduction, but Figure 2 does NOT prove it.

Finally, even if an association were causal for the industry as a whole, it does not follow that the association means an individual cow with higher production would necessarily be more at risk of, in this case, poor reproduction. What is true as a dynamic across a population does not necessarily extend to the individual. Consider a non-dairy example: It is clear that over the past several decades people in the U.S. have been more at risk of obesity and the attendant illnesses that go along with obesity. It is also true that we have become more affluent as a society over the same period. Finally, it is probably true that the two (affluence and obesity) are causally connected for the population as a whole. This does not mean, however, that a person winning the lottery would be expected to become obese. This non-dairy example makes sense to most people, but doesn’t stop many in the dairy industry from making the false logical leap that if reproduction has become poorer while production has increased in the population, it must be true that high producing cows must therefore suffer from poorer reproductive performance.

1. We know production has been increasing. Is there more disease?

There are plenty of data to show that over the past five decades production per cow has increased. What do the data tell us about disease rates? Before looking at the literature, some thought about the quality of the data is needed.

First, for most major clinical diseases, there is a remarkable paucity of quality data regarding the incidence of the disease. Much of the data are derived from university studies where the researcher collected data on a convenience sample of herds, typically university herds or cooperating local dairy farmers. These data may not represent a fair picture of what is happening across the industry. The only truly cross-sectional studies of the dairy industry across the U.S. are those conducted by the National Animal Health Monitoring System (NAHMS, USDA) and they are only just now conducting the second national round of dairy surveys. Even NAHMS must deal with disease data that are producer reported and subjective in many cases.

Second, the issue of who diagnoses the disease, what is the definition of the disease, and how reliably is disease recorded all have to be considered. For some diseases, e.g. retained placenta, there is a fair degree of agreement of what constitutes a case (retention past one day) and the diagnosis is fairly obvious. What about ketosis? Was each cow’s urine, milk, or blood checked each day in the post-partum period? What test was used and how sensitive and specific was the test? Who made the diagnosis? How ketotic must the cow be to count as a case? How reliably was it recorded? It is obvious that studies of the incidence of ketosis will be fraught with problems of inconsistency across even well conducted studies.

Why else might disease seem to be associated with production? There may be systematic bias introduced into the data, particularly by deliberate culling decisions. Low producing cows with repeated bouts of mastitis might be culled while high producing cows might be kept. After culling the mastitis low producers, a survey of the herd would show an association between high production and mastitis. Some perceived associations might be nothing more than faulty perception because we can observe larger populations of cows. A 4% death rate in a 50-cow dairy means that 2 cows die on the dairy each year. Those two deaths are probably easily “explained away” as unusual, exceptional events, not indicative of any particular causality. The same 4% death rate in a 1,000 cow dairy is 40 cows or nearly one each week. This apparent problem begs for an explanation, even if it may be a fabricated one. Finally, our record systems and our ability to capture accurate and complete data have been steadily improving. What appears to be an increased incidence may be nothing more than increased quality of data recording.

With these difficulties firmly in mind, it is still interesting to consider what we know about the rates of disease and the association or connection of disease and production over the past several decades. If there were evidence that the rate of a disease had increased during the period when production also increased, then one could form a hypothesis that increased production might lead to more of the disease. The association would not be proof, but could stimulate more reliable testing of the hypothesis by other sorts of studies.

Left Displaced Abomasum

Left displaced abomasum (LDA) is a fascinating problem in the dairy cow. Before the 1950s, the disease was simply not known. The 1916 USDA compendium of cattle disease lists most of the digestive disease we know today and a few we would not often diagnose: bloat (acute and chronic), feed overload, hardware (reticuloperitonitis), hair concretions, vomiting, indigestion or dyspepsia. Notably, the book makes no mention of displacement of the abomasum. While it is possible that veterinarians simply overlooked the disease, it seems implausible that no surgeon or pathologist cut open an ill or dead cow and discovered the abomasum was in the wrong place. It seems far more reasonable to conclude that the disease didn’t exist in 1916.

The first documented cases of LDA in the English literature were reported in England in 1950 (Ford 1950), followed swiftly by other reports (Begg 1950). Still, the disease was quite rare and prompted American reviewers of the condition in 1954 to open their paper with “Displacement of the abomasum is one of several obscure conditions that are of chief importance because they cloud and confuse diagnosis of digestive and metabolic diseases in cattle practice.” (Moore 1954) Less than a decade later, the authors of a case series of 80 cases opened their paper quite differently: “Perhaps one of the most remarkable features of dairy practice in the last decade has been the apparent marked rise in the incidence of displacement of the abomasum of the dairy cow. When first reported by Begg (1950) and Ford (1950) there was no indication that within a few years this condition would prove to be one of the most common surgical conditions of the bovine alimentary tract.” (Pinsent 1961) In some respects, the sudden appearance of the disease mimicked the introduction of a new infectious pathogen into a susceptible population.

Later surveys of the incidence of LDA in herds have shown a fairly wide variation in incidence. A Canadian survey of 32 Ontario herds reported an incidence of LDA of 1.2% (Dohoo 1982). An early state level NAHMS survey in Ohio reported an incidence of 8.4% (Ohio 1986, Miller), while similar state surveys in California (Hird 1986) and Michigan (Kaneene 1990) did not report LDA as a category (too few cases). A 1995 report from New York reported an incidence of 6.3% in 25 herds with 8,000 cows (Groehn 1995). The national 1996 NAHMS dairy study reported an incidence of 2.8% (Wells 1996). These were owner reported cases, so the incidence is probably a conservative estimate of the true rate.

While the rates may differ from one study or decade to the next, it is clear that LDA is now a common disease of dairy cows, a very different status than existed 50 years ago. Something has fundamentally changed that puts the cow at more risk of LDA, whether management (feeding, feeds, housing, etc.) or the cow herself.

Reproduction

General Performance

As Figure 2 shows, there appears to have been a distinct shift in reproductive performance over the past two decades in reported overall reproductive performance in dairy cows. The average days open in Minnesota dairy herds has increased from 120 to 170 days. Figure 3 shows the data for average conception rate in pregnant cows (MN DHIA). In the past two decades conception rate in pregnant cows has dropped 10 percent, from 58% to 48%. Some of these shifts may reflect changes in the behavior and policy of the producers themselves in terms of how long they are willing to breed open cows. Some of the change may reflect improvements in record keeping and more accurate reporting of the raw data through DHIA systems. As much as we would like to use these data to “prove” that there has been a fundamental shift in the basic efficiency with which cows become pregnant on dairies, we only have a correlation. An associated question is whether the incidence of specific reproductive diseases has also changed over time.

There is no doubt that as milk production increases, many dairy farmers will keep cows longer. As they have learned to focus on pregnancy instead of conception rate, they will also be more aggressive at breeding more cows, even if their “reported” reproductive performance looks worse. Ten years ago, many farms did not have on-farm software, and it was not uncommon for less-progressive farms to only report the cow’s final breeding to DHIA. Much of the apparent effect we see in reproductive performance might be nothing more than better data recording on better farms. Particularly in the past decade there has been an increase in the use of synchronized and timed breeding programs. Effects on first service conception may not reflect any real change in the cow’s basic physiology. As these practices have changed and we have fewer cows that “conceived” on their first breeding, it appearsthat first-service and overall conception rates are decreasing.

Metritis

With the data available, it is difficult to decide whether there has been a shift in the incidence of metritis. In a Canadian study in the late seventies, the reported incidence of metritis was 18.2% (Dohoo 1982). Early NAHMS state surveys reported an incidence of 35% (Ohio 1986, Miller) and 7% (California 1986, Hird). Such wide variation in reported incidence almost certainly reflects difference in case definition and reporting, not true differences in the disease itself. In a later New York survey, metritis incidence was 7.6% (Groehn 1995) and using a different set of dairies in 1998 4.2% (Groehn 1998).

Retained Placenta

The reported incidence of retained placenta shows much less variability than metritis. In the Canadian study, the incidence was 8.6% (Dohoo 1982). State NAHMS surveys reported 8.0% (Ohio 1986, Miller), 4.7% (California 1986, Hird). Surveys in New York showed 7.4% (Groehn 1995) and 9.5% (Groehn 1998). The National NAHMS survey reported 7.8% (NAHMS 1996). Surveying retained placenta rates has the advantage that the case definition is clear and the problem is easily seen, probably making the data more reliable than for other diseases. At least in the last two decades, it appears that the average rate has been fairly consistent at roughly 8 percent of cows calved. Thus there is fairly good evidence that as production has increased over the past two decades the incidence of retained placenta has not changed. Increasing production does not appear to increase the risk of retained placenta, at least as measured by considering the association between the two.

Mastitis

There are two dimensions to the question of whether the rate of mastitis infection has shifted over time as production has increased. The first dimension is the prevalence of subclinical mastitis (measured by average level of somatic cells {SCC}) and second is the incidence of clinical mastitis .

Somatic Cell Count

Figure 4 shows the herd average log somatic cell count for Minnesota dairies over the past decade (MN DHIA). There has been a distinct rise in average cell count for Minnesota dairies over that time. Figure 5 shows the average percent of cows in DHIA herds that were SCC positive (log SCC of 4 or greater), this time over a two-decade period (MN DHIA). The data fluctuate some, but there is a distinct upward trend that a greater proportion of cows are above the traditional cutoff for mastitis. Some of the observed change in these DHIA data might be explained by changes in the population of herds testing with DHIA. If the use of SCC testing gradually extended down into herds with relatively poorer management, then the total DHIA average would worsen even if no individual herd performed more poorly. This confounder could be avoided by following herds over time. Ott recently conducted a survey of herds in four regions (Upper Midwest, Mideast, Central and Southwest). In the survey of more than 15,000 herds that were followed from 1997 to 2001, the average SCC in that time period increased from 307,000 cells ml to 320,000 (Ott 2002). Despite efforts to improve milk quality, average levels of subclinical infection seem to be increasing.

Clinical Mastitis

Again, we need to understand the distinction between a biological cause for any apparent association and a management cause. One could hypothesize plausible biological reasons why clinical mastitis might be associated with higher production. High milk production likely causes higher flow rates during milking. Older milking equipment might be poorly adjusted to handle these flows and lower teat-end vacuums would occur. Fall-offs and slower milking might be more common in higher producing cows, possibly leading to teat-end lesions and both sub-clinical and clinical mastitis cases. High producing cows might genetically have more open teats, putting them at risk of more mastitis. On the other hand, in management terms, a low producing cow with a case of clinical mastitis is far more likely to be culled than a high producing cow with the same disease. When surveys are done, the low producing cows have left the herd and may not be included, and it appears that mastitis has “caused” high milk production. If an association is found between high production and mastitis, one still does not know which type of cause is operating.

Surveying the incidence of clinical mastitis suffers from all of the usual problems of case definition, testing strategy, and reporting errors. Published incidence rates vary quite widely. Canadian data from the late 1970s and early 1980s reported an incidence of 16.8% in the (Dohoo 1983). State NAHMS data reported 40% (Ohio 1986, Miller) and 30.3% (California 1986, Hird). The 1996 National NAHMS survey reported an incidence of 13.4%. The two New York studies reported 9.7% and 14.5% (Groehn 1995 and 1998, respectively). It is very uncertain whether the incidence of clinical mastitis has changed as production per cow in the industry has increased .

Milk Fever

Rare in first calf heifers, milk fever has long been assumed to be associated with milk production. The 1916 USDA Diseases of Cattle says: “It is the disease of cows that have been improved in the direction of early maturity, power of rapid fattening, or a heavy yield of milk, and hence it is characteristic of those having great appetites and extraordinary power of digestion.” Incidence rates published show a variety of levels. Since the disease is readily diagnosed, variations in incidence probably reflecting feeding management in the sample of herds studied more than anything else. In the 1970 veterinary textbook Bovine Medicine and Surgery, Kronfeld and Ramberg described the incidence as 3 to 7 percent of calvings (Kronfeld 1970). Canadian data reported an incidence of 2.9% as down cows and an additional 7.9% standing milk fevers (Dohoo 1982). California NAHMS survey reported 4.7% (Hirt 1986). New York data reported 1.6% (Groehn 1995) and 0.9% (Groehn 1998). National NAHMS reported 5.9% (NAHMS 1996). There is little compelling incidence data to suggest a significant increase in the incidence of milk fever as production has increased.

Ketosis

Ketosis is a disease particularly fraught with problems of case identification. The tests are inconsistently applied and vary in their sensitivity and specificity. Recording and reporting the disease is erratic as well. Kronfeld and Emery described an incidence of 2 to 20 percent of calvings in their chapter in a 1970 textbook (Kronfeld 1970). In the Canadian survey, an incidence of 7.4% was reported (Dohoo 1982). In the two New York studies, and incidence of 4.6% and 5.0% was reported (Groehn 1995 and 1998, respectively). Again, there is little compelling incidence data (however unreliable) to suggest a significant increase in the incidence of ketosis on dairies as production has increased.

General Synopsis of Studies

Using the data detailed above, it seems supportable to conclude that as average production per cow in the dairy industry has increased over the longer term, there has probably also been a decrease in reproductive efficiency and an increase in mastitis (at least subclinical mastitis). Since the 1950s, left displacement of the abomasum has become a prominent disease of the digestive tract, but it is more problematic whether the disease has become more common in more recent years than it was, for example, in the 1970s. While there are significant concerns about the quality and comparability of data from studies across time and between herds, there is less compelling data that the incidence of ketosis, milk fever, metritis, or retained placenta has increased in the past two decades as production has climbed.

In 1987, Dr. Hollis Erb did a thorough epidemiologic review of the available literature on this topic (Erb 1987). The summary from that paper says: “Epidemiologic evidence is presented in order to answer two questions. The first question is: ‘Does high milk production put a cow at increased risk of disease?’ The answer to this question seems to be ‘maybe’ for milk fever, but ‘no’ for most other common diseases (veterinary assisted dystocia, retained placenta, metritis, cystic ovary, ketosis, left displaced abomasum and mastitis). The second question is: ‘Is low milk production a consequence of disease?’ For most diseases the answer is a cautious ‘yes’.” Dr. Erb’s conclusion is supported in general by studies reported after her review.

2. Cohort Studies

Rather than just looking for associations between production and disease or changes in incidence over time as production has increased, there is an alternative and stronger study approach to the question of production and disease. In this approach, a cohort of cows with known levels of production can be identified at the beginning of the study period and then followed over time, recording cases of clinical disease. By analyzing disease incidence at different levels of production, a more direct measure of the effect of production on disease is possible.

In one such study, 2,875 lactation records in 32 commercial herds in Ontario from 1979 to 1981 were considered. Production levels were set by previous lactation Breed Class Average to adjust for differences in ages of cows and further adjusted for the herd of origin’s level of production to remove the confounding effect of the herd’s level of production and management (Dohoo 1984). The authors concluded that “the level of milk production was not related significantly to the risk of any of the common disease conditions” except for an association between previous production and milk fever.” (i.e. no effect on dystocia, retained placenta, displaced abomasum, ketosis, metritis, clinical mastitis, among others). The “study found no association between previous milk production and the risk of reproductive diseases.” In summarizing other previous studies of the effect of production on reproduction, the authors concluded that “most studies do not indicate major associations between previous production and reproductive diseases and either no association, or a small negative association between early lactation milk production and reproductive performance.”

In a study of 15,320 Holstein cows in 26 herds in New York between 1990 and 1993, associations between milk yield, days open and days to first breeding were investigated (Eicker 1996). In this study, milk production in the first 60 days of lactation were used as the measure of production and results were adjusted for season, herd, and parity. Cows in the highest 20% of production had slightly lower conception rate than cows in the lowest 20% of production. Regarding reproduction, the authors concluded that “these results indicate that conception and insemination might be influenced by factors related to management (e.g. culling) and to the cow (e.g. disease history) but that increased milk yield plays a very minor role.”

In a separate study of 8,070 cows in 25 New York dairies between 1990 and 1993, 305-day milk yield in the previous lactation was studied in association with disease incidence (Groehn 1995). Higher milk production was not associated with increases in the risk of retained placenta, metritis, ovarian cyst, milk fever, ketosis, or abomasal displacement. There was a small association between increased milk production and clinical mastitis. In this study the authors did not find the small effect on the risk of milk fever found in the Canadian study, but did find an association between production and clinical mastitis not found in the Canadian study.

It may clarify the impact of production on clinical mastitis by quantifying the effect measured in the New York study. Based on the results of the study, increasing production from about 22,300 to about 24,700 pounds per lactation (an increase of 2,400 pounds of milk) increased the risk of clinical mastitis 15% (increases the baseline risk of clinical mastitis by a factor of 1.15). If the baseline risk of clinical mastitis at the lower production level were 20% of the herd per year, then the additional 2,400 pounds of milk would increase the risk of clinical mastitis to 23% of the herd per year, an increased risk of 3%. Thus if a 100 cow herd increased production by this amount in every cow, they could expect to see about an additional 3 cases of mastitis per year.

3. Disease Effects of Increasing Production Using rbST

Since the approval of recombinant bovine somatotropin (rbST) use in dairy cows in the United States (POSILAC 1 STEP , Monsanto), there has been another “natural experiment” that can be used to consider whether increased production increases disease. For many diseases, POSILAC and the subsequent increased milk production cannot be a risk factor because the disease occurs in early lactation, before POSILAC is administered. Thus the increased production from POSILAC use cannot increase the risk of milk fever, ketosis, etc. unless a carry-over effect from the previous lactation exists. No such negative carry-over effects of POSILAC use have been demonstrated. Interestingly, use of POSILAC does seem to reduce the incidence of ketosis in cows at the start of the subsequent lactation (Dohoo 1999).

For other health problems occurring during the time of administration of POSILAC, extensive pre-approval and post-approval field studies have shown some negative health impact from the use of POSILAC and the subsequent increased milk production. The Post Approval Monitoring Program (PAMP) studied 1,213 cows in 28 herds across the United States (Collier 2001). In a separate review as part of the Canadian governments review of POSILAC for use in Canada, a team of veterinary scientists developed a summary of all of then available data on the impact of rbST on health (Dohoo 1999). The review included as many as 18 separate studies,depending on the parameter being considered. The review included data not just for the product being considered, but other experimental formulations of rbST as well. A summary of the conclusions from these studies is:

Reproduction

The Canadian review concluded that rbST use increased average days open (by about 5 days) and saw a trend of increases in cystic ovaries. Their review concluded that rbST use resulted in an increase in percent of cows not becoming pregnant. There was no conclusive evidence that rbST use increased the rate of twinning or abortion (Dohoo 1999). Post approval studies in the U.S. have found that POSILAC supplemented cows are at no increased risk of twinning, have no difference in gestation length, successful calving rate, or percent pregnant, and could not establish a statistically significant increase in days open (Collier 2001). The data on impacts on reproduction are variable and contradictory, but if there is an impact of production on reproduction, it seem likely to be small compared to the range of variability between farms for reproductive performance.

Mastitis

The Canadian panel concluded: “Use of rbST increased the risk of clinical mastitis by approximately 25%.” It is interesting to compare this estimate of the impact of rbST (and production increases) to the New York study referenced above that showed an increase of 15% in cows making 2,400 pounds more milk. If rbST were used from day 60 to day 300 in lactation and production increased 10 pounds of milk per day, then total production increase would be the same 2,400 pounds. It appears that the magnitude of increased risk of mastitis with increased production (whether from rbST or by other means) is the same order of magnitude. Using the same 100 cow herd example above and a baseline incidence of clinical mastitis of 20%, then if the rate increased by 25%, then the new clinical mastitis rate would rise to 25% of cows or 5 additional case per year (.20 * 1.25 = .25). Differing from the Canadian summary of earlier studies, the post-approval study concluded that “Supplementation of cows with POSILAC had no effect on total mastitis cases, total days of mastitis, duration of mastitis, or the odds ratio of a cow to develop mastitis” (Collier 2001). Viewed broadly, the data from rbST studies and summaries would seem to indicate that increased production may put cows at higher risk of clinical mastitis, but at the same level as the effect of any increase in production.

Other Health Effects

Cows supplemented with rbST have a small increase in the amount of musculoskeletal illness, including foot disorders and disorders of the joints, particularly the hock.

Summary

This paper has looked through three different windows at the question of whether increased production puts a cow at greater risk of disease: changes in rates of disease across the decades as production increased, cohort studies, and the impact of increase production from rbST use. While there are some cases where it is certain that the risk of illness has increased across the industry as whole (e.g. LDA), it is much more problematic to conclude that the incidence of many of the more common diseases has actually increased in the past decades. It is clear that the dairy industry of today is not the same as the industry of 1950, but at the same time there is little evidence that the industry is suffering from some expanded epidemic of disease as a result of consistent gains in production per cow. Results of cohort studies reach similar conclusions. With the exception of clinical mastitis, there is little evidence that higher producing cows are at risk for increased disease. Finally, rbST and increasing milk production has little impact on cow health beyond clinical mastitis and minor and variable effects on reproduction and locomotor disease.

In sum, the paradox seems to be largely one of perception, not reality. It is true that healthy, comfortable, well-fed cows make more milk. The technology of dairy farming and the genetics of the cow have changed dramatically in the past decades. Production has increased steadily. In some cases, rates of diseases suffered by dairy cows have changed as well when viewed across the industry as a whole. What has happened at an industry level translates very poorly (if at all) to the individual cow. With a few exceptions, there is little reason to believe that improving production in an individual cow by improving the herd’s management will increase her risk of health problems.

The difference is the distinction between what happens in the aggregate as the dairy industry’s technology has changed and how those changes will affect an individual cow: e.g. feeding, housing, genetics, health management, etc. These technology changes have created the necessary conditions to increase the risk of specific diseases. The risks increase across herds that adopt these technologies. The risk is not necessarily greater for high producing cows within a herd, in fact there is at least an argument that these cows are healthier (less likely to be broken) and therefore able to be high producers.

POSILAC 1 STEP  is a registered trademark of Monsanto Technology LLC.

References

Begg, H. Diseases of the stomach of the adult ruminant. Vet. Rec. 62:797-803, 1950.

Collier RJ, Byatt JC, Denham SC, et al. Effects of sustained release bovine somatotropin (Sometribove) on animal health in commercial dairy herds. J. Dairy Sci. 84:1098-1108, 2001.

Dohoo IR, Martin SW, Meek AH, Sandals WCD. Disease, production and culling in Holstein-Friesian cows: I. The data. Prevent Vet. Med. 1:321-334, 1982.

Dohoo IR, Martin SW. Disease, production and culling in Holstein-Friesian cows: III. Disease and production as determinants of disease. Prevent. Vet. Med. 2:671-690, 1984.

Dohoo I, DesCouteaux L, Dowling P, et al. Report of the Canadian Veterinary Medical Association expert panel on rbST, 1999. Accessed August, 2002.

URL: http://www.hc-sc.gc.ca/english/protection/rbst/animals/app05.htm

Eicker SW, Groehn YT, Hertl JA. The association between cumulative milk yield, days open and days to first breeding in New York Holstein cows. J. Dairy Sci. 79:235-241, 1996.

Erb, HN. Interrelationships among production and clinical disease in dairy cattle: a review. Can. Vet. J. 28:326-329, 1987.

Ford EJH. A case of displacement of the bovine abomasum. Vet. Rec. 49:763-764, 1950.

Groehn YT, Eicker SW, Hertl JA. The association between previous 305-day milk yield and disease in New York State dairy cows. J. Dairy Sci. 78:1693-1702, 1995.

Groehn YT, Eicker SW, Ducrocq V and Hertl JA. Effect of disease on the culling of Holstein dairy cows in New York State. J. Dairy Sci. 81:966-978, 1998.

Hird DW and Sischo WM. Studies in data management and data reporting for California’s Round Two of the National Animal Health Monitoring System 1986-1987. Report submitted to the Veterinary Services, Animal Plant Health Inspection Service, USDA, February 1988.

Kaneene JB, Hurd HS, Miller R, and Lafi SQ. Michigan National Animal Health Monitoring System Round Two Final Report 1988-1989. USDA APHIS VS, February 1990.

Kronfeld DS, Ramberg CF. Parturient Paresis. Chapter in Bovine Medicine and Surgery, pp 382-394. American Veterinary Publications, 1970.

Kronfeld DS, Emery RS. Acetonemia. Chapter in Bovine Medicine and Surgery, pp 350-376. American Veterinary Publications, 1970.

Miller GY. Evaluation of economic and epidemiologic data from the National Animal Disease Detection System (NADDS). Final contract report, USDA #12-16-93-182, October 1987.

MN DHIA. Minnesota DHIA summary statistics. Compiled in 2002 by Barbara Schmidt, Sales and Marketing Specialist, Minnesota DHIA, Buffalo, Minnesota.

Moore GR, Riley WF, Westcott RW and Conner GH. Displacement of the bovine abomasum. Vet. Med. pp 49-51, February 1954.

Ott SL. Five year trend shows rise in U.S. bulk tank somatic cell counts. Udder Topics: National Mastitis Council newsletter 25:1, June-July 2002.

Pinsent PJN, Neal PA, Ritchie HE. Displacement of the bovine abomasum: a review of 80 clinical cases. Vet. Rec. 73:729-735, 1961.

United State Department of Agriculture, Bureau of Animal Industry. Special Report of the Diseases of Cattle, Government Printing Office, 1916.

USDA Animal Plant Health Inspection Service Veterinary Services. 1996 National Dairy Study report: NAHMS 1996. May 1996, accessed via URLhttp://extranet.aphis.usda.gov/vs/ceah/cahm/Dairy_Cattle/dairy.htm

Figure 1: Minnesota DHIA: production per cow and percent of herd culled per year