Coxiella burnetii infection in humans: to what extent do cattle in infected areas free from small ruminants play a role?

Domestic ruminants (cattle, goats and sheep) are considered to be the main reservoirs for human Coxiella burnetii infection. However, there is still a need to assess the specific contribution of cattle. Indeed, most seroprevalence studies in humans were carried out in areas comprising both cattle and small ruminants, the latter being systematically implicated in human Q fever outbreaks. Therefore, we conducted a cross-sectional study in areas where C. burnetii infection in cattle was endemic, where the density of cattle and small ruminant farms were respectively high and very low. The aim was to estimate the seroprevalence rates among two occupational (cattle farmers and livestock veterinarians), and one non-occupational (general adult population) risk groups. Sera were collected in 176 cattle farmers, 45 veterinarians and 347 blood donors, and tested for phase I and II antibodies using immunofluorescence assay. Seroprevalence rates were 56.3% among cattle farmers, 88.9% among veterinarians and 12.7% among blood donors. This suggests that a specific risk for acquiring C. burnetii infection from cattle in endemically infected areas exists, mainly for occupational risk groups, but also for the general population. Further research is needed to identify risk factors for C. burnetii infection in humans in such areas.

Does feeding extruded linseed to dairy cows improve reproductive performance in dairy herds? An observational study.

Feeding n-3 fatty acids (FA) is often cited as a promising strategy to tackle impaired reproduction in dairy cows. However, the scientific literature shows conflicting results that may be explained by the nature of n-3 FA used, the amount supplemented and the timing of supplementation. In addition, designing a proper experimental design to study n-3 FA and reproduction is subjected to other difficulties such as the choice of the control diet or gaining enough statistical power. The objective of this retrospective observational study was to quantify the average effects of supplementing extruded linseed (EL), a feed rich in α-linolenic acid, to dairy cows on reproductive performances under field conditions in French commercial farms. Exposure measurement to EL feeding was particularly challenging as exact cow diets are not traced in farms. Therefore, to investigate the potential dose-effect relationship, we defined a proxy of EL intake per day by using deliveries of EL based feeds from 22 companies in the study period 2008-2015 in France. An artificial insemination (AI) was considered exposed only if the cow was supplemented with EL from the calving until 17 days after AI. Based on recommendations for EL use on the field, 4 exposures classes were created: [1-50] (n = 14,126 AIs), [50-300] (n = 88,261 AIs), [300-600] (n = 66,136 AIs), and [600-1500] (n = 28,287 AIs) g/cow/d. The reference population was composed of cows that did not receive any EL between calving until 17 days after AI within herds that were supplied, but not continuously during the study period (n = 226,795 AIs). Mean daily EL intake in exposed population was 337 g/cow/d (±239.4). Reproductive performance was studied on 423,605 AIs from 1096 herds and 158,125 cows using Cox models for days to first AI and days to conception, and logistic regression models for risk of return-to-service, adjusted for factors likely to influence the reproductive performance and for a herd random effect. Risk of return-to-service between 18 and 78 days after first and second AI did not differ between exposed and reference populations, Nevertheless, the effect on the days to first AI was higher with the lowest EL intake (HR: 1.14; 95% CI: 1.11, 1.17) than with higher EL intake levels (HR ranging from 1.06 to 1.07; 95% CI: 1.04, 1.09). Similarly, for the effect on the time from calving to conception from the lowest EL intake (HR: 1.19; 95% CI: 1.15, 1.23) compared to the higher EL intake levels (HR ranging from 1.08 to 1.11; 95% CI: 1.06, 1.14). This original large-scale epidemiological study provides new insights into the effects of feeding EL at a commercially sustainable level to dairy cows.

Inpatient satisfaction with medical information received from caregivers: an observational study on the effect of social deprivation.

BACKGROUND: The main objective of this study was to explore the relationships between inpatients' social differentiation and satisfaction with the medical information delivered by caregivers. METHODS: In four departments of a teaching hospital, patients were enrolled as well as their attending physician and one of the nurses assigned to them. Structured survey questionnaires were administered face-to-face to patients and caregivers. Patients were asked to rate their satisfaction with the medical information received, the quality and duration of the interactions with the caregivers, and their experience regarding their involvement in medical decision-making. Caregivers were asked to rate their perception of the patients' social position and involvement in medical decision-making. Social deprivation was assessed using the EPICES score in particular. The statistical analysis was mainly descriptive and completed by a structural equation model. RESULTS: A sample of 255 patients, 221 pairs of patient-physician and 235 pairs of patient-nurse were considered. One third of the patients (32.7%) were identified as socially deprived. They were significantly less satisfied with the information they received on their health status or their treatment; 56.7% of patients thought that they received sufficient explanations without having to ask. This proportion was significantly lower in socially deprived patients (42.3%) compared to not deprived patients (63.6%, p < 0.01). Patients' reported involvement in medical decision-making was significantly lower for socially deprived patients (75.0% vs 89.0%, p < 0.001). The structural equation model showed that the main determinant of patients' satisfaction regarding medical information was their perceived involvement in informed medical decision-making (CFI = 0.998, RMSEA = 0.022). CONCLUSIONS: These findings suggest that physicians and nurses need training on communication targeted towards vulnerable patients, in order to improve the accessibility of medical information, and thus to reduce health inequalities.

The Effect of Wind on Coxiella burnetii Transmission Between Cattle Herds: a Mechanistic Approach.

There is a consensus that wind plays a key role in the transmission of Coxiella burnetii, the causative agent of Q fever, between ruminants and from ruminants to humans. However, no observational study so far has focused on the mechanisms associated with this airborne transmission. This study applied a mechanistic epidemiological approach to investigate the processes underlying the wind effect and to assess its influence on the risk for a dairy herd to become C. burnetii infected. Ninety-five dairy cattle herds located in the Finistère department (western France) were subjected to samplings of bulk tank milk and indoor dust every 4 months over a 1-year period to determine their C. burnetii status using PCR tests. A total of 27 incident herd-periods (negative-tested on both PCR tests and becoming positive-tested at least once at the subsequent sampling time) and 71 negative herd-periods were retained for analysis. Using logistic regression, we assessed the effect of (i) the cumulated number of bacteria in herds located under the main wind direction and (ii) the mean wind speed in this area, on a given herd's risk of becoming incident. Compared to herds in areas with low wind speed (≤5.5 m/s), the risk was significantly higher (OR = 3.7) in herds in areas with high wind speed (>5.5 m/s) and high bacterial load (>10), whereas it was not significantly different from unity in other situations. In agreement with our assumptions, C. burnetii transmission to a previously infection-free herd occurs only when (i) the wind transporting from infected sources and (ii) the load in the contaminated particles/aerosols generated are high enough to act jointly.

Q fever infection in dairy cattle herds: increased risk with high wind speed and low precipitation.

Ruminants are considered the main reservoir for transmission of Coxiella burnetii (Cb) to humans. The implementation of effective control measures against Cb in ruminants requires knowledge about potential risk factors. The objectives of this study were (i) to describe the spatial distribution of Q fever-infected dairy cattle herds in Sweden, (ii) to quantify the respective contributions of wind and animal movements on the risk of infection, while accounting for other sources of variation, and (iii) to investigate the possible protective effect of precipitation. A total of 1537 bulk milk samples were collected and tested for presence of Cb antibodies. The prevalence of test-positive herds was higher in the south of Sweden. For herds located in areas with high wind speed, open landscape, high animal densities and high temperature, the risk of being infected reached very high values. Because these factors are difficult to control, vaccination could be an appropriate control measure in these areas. Finally, the cumulated precipitation over 1 year was identified as a protective factor.

Decrease in milk yield associated with exposure to bluetongue virus serotype 8 in cattle herds.

Decreased milk yield and reduced fertility are the primary consequences of infection by bluetongue virus serotype 8 (BTV-8). These effects must be quantified to fully assess the economic benefit of vaccination. This can be estimated by measuring the effect of BTV-8 exposure on milk yield and fertility for all cows belonging to an infected herd. The objectives of this study were (1) to quantify the mean effect of exposure to BTV-8 on milk yield following natural challenge for cows in herds previously naïve, (2) to determine the duration of reduced milk yield before and after the date disease was first detected in the herd to estimate the cumulative loss of milk yield during this period, and (3) to evaluate the influence of the proportion of infected neighboring herds on the reduction in milk yield following exposure to BTV-8. The effects of exposure to BTV-8 during the French outbreak of 2007 were assessed using mixed linear models, which allow adjustment for factors known to influence milk yield. Exposure to BTV-8 was associated with a sharp decrease in milk yield over a period of 6 mo (2 mo before to 4 mo after the reported date of disease detection in the herd). The cumulative loss of milk yield was more than 3% of annual production. The relatively earlier reduction in milk yield in infected herds detected later in the outbreak period suggests that detection of clinical signs was delayed in these herds. Finally, the greatest decrease in milk yield was observed in herds detected early during the outbreak period and located in areas with the highest disease incidence. This may be due to a greater within-herd incidence or to a greater amount of virus injected by midges to individual cows in these herds.

Increase in the occurrence of abortions associated with exposure to the Bluetongue virus serotype 8 in naïve dairy herds.

The transplacental transmission capacity demonstrated for Bluetongue virus serotype 8 (BTV-8) in cattle probably is associated with an increased occurrence of abortions. The objectives of this study were to quantify the effect of BTV-8 exposure on the occurrence of abortions in previously naive dairy cow herds under natural infection conditions, and to determine a possible risk period during pregnancy associated with this increase. Two criteria were considered in order to estimate the occurrence of abortion: late return-to-service after a first artificial insemination (AI), and short gestations. A late return-to-service was defined as a return taking place 90 to 200 days after a first AI. These criteria were compared between cows in herds exposed during the 2007 epizootic in France and cows in herds that were not exposed. To determine the risk period during a pregnancy, variations in the occurrence of abortions were quantified according to the stage of the pregnancy during which the exposure took place. Survival analyses were used to estimate the risk of increased occurrence of abortion associated with BTV-8 exposure, adjusted by the principal factors known to influence the risk of abortion. Exposure to the BTV-8 virus under natural conditions in previously naive dairy herds notified after clinical suspicion during the 2007 epizootic was associated with an increase in the occurrence of abortions, regardless of the stage of pregnancy. The at-risk gestation period depended on the criteria used to detect abortions. The mean effect of BTV-8 exposure in the ensemble of detected outbreaks corresponded to an increase of 6.7% in late return-to-service. BTV-8 exposure during the first 3 mo of gestation was associated with a 15% increase in late return-to-service for cows with no return-to-service at 90 days, while this increase was 6% for exposure starting from the third month of gestation (in outbreaks detected in September). BTV-8 exposure from the third month of gestation was associated with a 1.9% increase of short gestations. The effect of exposure was more pronounced for outbreaks detected early in the epizootic compared with those detected later.

Quantification and at-risk period of decreased fertility associated with exposure to bluetongue virus serotype 8 in naïve dairy herds.

The detrimental effect of bluetongue virus serotype 8 (BTV-8) on fertility was quantified in seroconverting cows. Although the effect on individual cows provides information regarding the potential biological burden of infection, losses at a herd level are also dependent on the proportion of infected cows within the herd. The objectives of this study were to quantify the average effect of BTV-8 exposure in field conditions on the fertility of dairy cows in previously naïve herds, and to determine the at-risk period of decreased fertility related to the date of detection of the disease in the herd. The effect of BTV-8 exposure on fertility was assessed using the 90-d-return-to-service rates after the first artificial insemination (AI) calculated for cows in exposed herds (during the 2007 epizootic in France) and compared with that for cows in unexposed herds. Only herds with a confirmed detection that were reported after clinical suspicion were included. To determine the at-risk period of decreased fertility, variations of fertility in exposed herds were quantified according to the time interval between the date of AI for individual cows and the date that disease was detected in the herd. Survival analyses were used to assess the risk of decreased fertility associated with BTV-8 exposure, adjusting for the main factors known to influence fertility. The episode at risk for decreased fertility depended on the month of disease detection in the herd. For herds detected early in the epizootic, fertility was decreased for cows inseminated from 1 mo before to 1 mo after the date of disease detection in the herd. Depending on time interval between the date of AI of cows and the date of detection in the herd, the increase of return-to-service rate associated with BTV-8 exposure varied from 8 to 21 percentage points of 90-d return to service. The episode of decreased fertility is likely due to a combination of the effect of the infection at different stages of conception and early pregnancy and the delayed exposure of cows due to the spreading of the virus within herds. For herds detected during the second half of the epizootic, fertility was decreased for cows inseminated more than 2 mo before detection, which suggests a delay in the detection of clinical signs following virus introduction in the herd. No correlation was observed between the effect of BTV-8 exposure on fertility and the incidence of BTV-8 in the local geographical area. Given the duration of the period that cows were at risk for decreased fertility and the magnitude of the effect, the average BTV-8 exposure in naïve herds led to major losses.

Shedding and serological patterns of dairy cows following abortions associated with Coxiella burnetii DNA detection.

To describe both shedding and serological patterns following abortions detected as being associated with Coxiella burnetii (Cb), 24 cows experiencing an abortion due to Cb were followed over a one month period. Samples taken on the day of abortion (D0) were followed 3-fold by weekly samplings from day 14 (D14) to D28 after the abortion. Milk and vaginal mucus were collected at each weekly sampling and tested using real-time PCR while blood samples were collected 2-fold on D21 and D28 and tested using ELISA. We found a very short duration of C. burnetii shedding in vaginal mucus after abortion, highlighting the need to collect samples as rapidly as possible following an abortion to avoid false negative results. In contrast with previous results, concomitancy of vaginal and mucus shedding was frequent, especially for cows shedding a high bacterial load on DO leading to the hypothesis that the clinical onset of the infection influences the modalities of Cb shedding. Lastly, serological results indicating a lack of sensitivity to detect Cb shedder cows (especially for cows for which Ct values were high) suggest that ELISA is not a useful tool to diagnose abortions at the individual level.

Relationship between the level of antibodies in bulk tank milk and the within-herd seroprevalence of Coxiella burnetii in cows.

The relationship between the level of antibodies in bulk tank milk (BTM) and the within-herd seroprevalence of Coxiella burnetii in cows was assessed. Blood from milking cows and BTM were sampled in 55 infected herds and tested using commercial enzyme-linked immunosorbent assay (ELISA) kits. The relationship between antibody levels and WHP, which was quantified using a general linear model, was only moderate (R(2)=0·15). Nevertheless, the lowest antibody level in BTM was associated with the lowest mean within-herd prevalence. The present finding indicates that ELISA applied to BTM could identify infected herds with quite low within-herd seroprevalence. For such herds, the vaccination of dairy cows as well as nulliparous heifers using a phase I vaccine could effectively prevent C. burnetii shedding.

Estimated prevalence of Mycobacterium avium subspecies paratuberculosis infection in herds of dairy goats in France.

A study of the prevalence of Mycobacterium avium subspecies paratuberculosis (MAP) infection in dairy goats was carried out in France, using a serological ELISA. One hundred and five herds were randomly selected in different regions, and all the goats aged six months or older (11,847 goats) were sampled. An estimation had previously been made of the accuracy of the ELISA, and the Rogan-Gladen estimator was used to convert apparent prevalence values into estimates of true prevalence values. Given that sensitivity and specificity at herd level are functions of sample size, and because the sizes of the herds varied considerably, the true prevalence of infection at herd level was estimated using a weighted average depending on the herd size stratum. At the level of the individual animal, the total true prevalence was estimated by applying a weighted average in accordance with the regional distribution of the animals in the target population. At herd level, apparent prevalence and estimated true prevalence were 55.2 and 62.9 per cent, respectively; at individual animal level they were 2.9 and 6.6 per cent, respectively. In herds that were seropositive for MAP, the average within-herd apparent prevalence was 5.9 per cent and the estimated average within-herd true prevalence was 11.1 per cent.

Associations between bovine coronavirus and bovine respiratory syncytial virus infections and animal performance in Swedish dairy herds.

To assess the economic impact of bovine coronavirus (BCV) and bovine respiratory syncytial virus (BRSV) infections, accurate estimates of their associated effects on animal performance are needed. This study aimed to quantify the variation in individual test-day milk yield and somatic cell count, risk of reproductive failure after first service of dairy cows, and risk of death of calves and heifers according to the BCV and BRSV status of the herd. Three types of status were defined for BCV and BRSV infections, based on 1) the dynamics over a 7-mo period of BCV- and BRSV-specific antibody levels in pooled milk of primiparous cows; 2) the possible occurrence of presumably BCV- and BRSV-related clinical outbreaks; and 3) the combination of both pieces of information. A total of 36,184 test days, 2,716 cows with a first service, and 4,104 calves and heifers in 65 Swedish herds were included in the analyses. Animal performance associated with BCV and BRSV infections was quantified using hierarchical mixed generalized and survival models, after adjustment for covariates known to influence the performance under study. A significant reduction in milk yield was observed for cows in presumably BRSV recently infected herds, as well as in herds having a presumably BRSV-related clinical outbreak (of 0.57 and 0.91 kg/d, respectively), compared with cows in presumably infection-free herds. There was also a significant increase in somatic cell count (of 12,000 cells/mL) for cows located in herds with a BRSV outbreak. The risk of failure after first service, as well as the risk of death in calf and heifer, was numerically higher in BRSV-infected herds, although this was not statistically significant. In contrast, BCV infection herd status, as defined in the present study, was not significantly associated with any production losses in animals from infected herds compared with those in infection-free herds.

Higher non-return rate associated with Mycobacterium avium subspecies paratuberculosis infection at early stage in Holstein dairy cows.

The effects of infection by Mycobacterium avium paratuberculosis (Map) on dairy cows are poorly documented and quite controversial. This retrospective study aimed at quantifying the variation in non-return to service of Holstein dairy cows according to their Map-infection status. Three different statuses were defined based on both individual and herd tests results: ELISA positive cow, all tests negative cow in a negative herd and all tests negative cow in a positive herd. Whatever the age at Map testing, the status was attributed to a cow from its first lactation onwards. Non-return to service was determined at 200 days after first and second services. The study was performed from 1999 to 2007 on 185,950 AI from 48,914 cows in early stage of the infection in 1069 herds by logistic regression controlling for known factors influencing non-return rate. Non-return rate was higher for infected cows compared to negative cows from negative herds (RR of 1.10 or +3.9 points of % of non-return rate). The effect was significant for parities 1 and 2 (RR of 1.11 and 1.12, respectively) but not for higher ones. This effect was lower when comparing positive cows to negative cows in the same herds but relative risks were still above 1. The hypothesis that the effect of Map on non-return depends upon the stage of infection is formulated.

Assessing the within-herd prevalence of Coxiella burnetii milk-shedder cows using a real-time PCR applied to bulk tank milk.

Thirty-seven bulk tank milk (BTM) and individual milk samples of all contributing cows were tested for Coxiella burnetii detection by a real-time PCR assay and used to assess the relationship between the BTM PCR-response and (i) the within-herd prevalence of milk-shedder cows and (ii) the proportion of heavy milk-shedder cows. The within-herd prevalence of milk-shedder cows (i) was found to be significantly higher in herds with a positive BTM and (ii) increased significantly with the estimated titre in Coxiella burnetii obtained in positive BTM. The proportion of heavy milk-shedder cows among the milk-shedder cows increased significantly with an increased estimated titre in Coxiella burnetii in positive BTM. Therefore, a real-time PCR assay applied to BTM samples collected repeatedly over time appears to be a valuable tool to assess on a larger scale the status of herds towards Coxiella shedding, and to evaluate the efficiency of control actions aimed at controlling and/or preventing Coxiella shedding in dairy herds.

Effects of sire and dam genotype for complex vertebral malformation (CVM) on risk of return-to-service in Holstein dairy cows and heifers.

The emergence of a new hereditary disease, called "Complex vertebral malformation" (CVM), has been described in Holstein cattle population. This paper studied the incidence of mating concerning Holstein dairy cattle with CVM in Brittany and the possible influences of CVM status for sires and dams on return-to-service at different intervals post-service in cows and heifers. It was carried out based on a set of data for first and second inseminations between 1998 and 2001 in cows (n=530,538) and heifers (n=248,140). Incidence of matings between CVM gene carriers, between a carrier bull and a non-carrier cow and between a non-carrier bull and a carrier cow were estimated to be 1.4, 10.6 and 9%, respectively (1.1, 9.3 and 10.9%, respectively in heifers). Compared to CVM-free mating, the relative risk of return-to-service was increased when the sire was a CVM carrier and the dam was at risk of being a carrier, especially for late return (>25 days post-service). When the sire alone was a carrier, the relative risk was increased whenever the return occurred at a low but significant level. Following mating between a non-carrier bull and a dam at risk of being a carrier, the risk of return-to-service decreased at a low but significant level, whatever the interval in the heifers but only for 19-25 days intervals in the cows.

Bulk milk somatic cell counts and bovine viral diarrhoea virus (BVDV) infection in 7252 dairy herds in Brittany (western France).

Variations in test-day bulk milk somatic cell count (BMSCC) according to bovine viral diarrhoea virus (BVDV) infection status of the herd were quantified. Infection statuses were defined, based on bulk-milk antibodies measured twice 4 months apart. On each test-day, BMSCC was calculated as the weighted mean of the individual cow-level somatic cell counts. A total of 97,985 test-days in 7252 herds were considered. The effect of the BVDV-infection on BMSCC was assessed using mixed linear models, controlling for herd (random), proportion of primiparous cows and average days in milk on test-day. Considering test-days contemporary to the BVDV-status definition-period, the increase in BMSCC was 12,000 (P<0.001), 27,000 (P<0.001) and 6000 (P=0.97) cells/ml in past-infected recently recovered, past-but-still-infected and recently infected herds respectively, compared with not recently infected herds. A carry-over effect (at least 1 year) of BVDV-infection on BMSCC was also evidenced in past-infected recently recovered, past-but-still-infected and recently infected herds.

Effect of quinolone treatment on selection and persistence of quinolone-resistant Escherichia coli in swine faecal flora.

AIMS: To study the effect of oral administration of a quinolone on emergence of resistance in an indicator bacterial species from faecal flora. METHODS AND RESULTS: Quinolone resistance was studied in Escherichia coli obtained from the faecal contents of pigs housed in nine commercial farrow-to-finish herds in France after administration of flumequine to sows. The percentage of quinolone-resistant E. coli increased in the faeces of sows after administration of flumequine (mean 21.78% at day 7 vs 6.42% before treatment for nalidixic acid) and then decreased (mean 12.6 and 10.4 at days 30 and 60, respectively for nalidixic acid), being not significantly different from initial values 1 month post-treatment. In young pigs, the proportion of resistant strains was lower and decreased over rearing period. Moreover, changes over time of both total E. coli and the proportion of resistant bacteria exhibited great inter-individual variability. CONCLUSIONS: Restoration of susceptible faecal flora occurred within 2 months after flumequine treatment. SIGNIFICANCE AND IMPACT OF THE STUDY: Effect of flumequine treatment of sows on the quinolone resistance of faecal E. coli of both sows and their progeny is noticeable but transitory.

Large scale assessment of the effect associated with bovine viral diarrhoea virus infection on fertility of dairy cows in 6149 dairy herds in Brittany (Western France).

This study aimed at quantifying on a large scale the risk of both three-week-return-to-service and late-return-to-service for dairy cows according to the BVDV-infection status of the herds. Three-week- and late-returns-to-service were defined as a repeat service occurring respectively from 19 to 25 days and after 25 days following a first or a second service. The level of BVDV-specific-antibodies in bulk tank milk were measured four times at four months intervals by a blocking ELISA test to define five BVDV-infection-herd-statuses: (i) presumed non-infected herds for a long time, (ii) presumed not-recently-infected herds, (iii) presumed past-infected-recently-recovered herds, (iv) presumed past-steadily-infected herds, (v) presumed recently-infected herds. A total of 150,854 AIs from 122,697 cows in 6149 herds was included in the analysis. The risk of return-to-service was assessed using logistic regression and survival analysis models. The BVDV-infection-herd-status was not significantly associated with the risk of 3-week-return-to-service, but significantly with that of late-return-to-service. Cows in herds presumed past-infected-recently-recovered, past-steadily-infected, or recently-infected had a significantly higher risk of late-return-to-service (relative risk of 1.03, 1.11, 1.12, respectively) compared with cows in herds presumed not-recently-infected. Risk of late-return-to-service was not significantly different in cows from herds presumed non-infected for a long time or not-recently-infected. In conclusion, BVDV-infection was found to mainly increase the risk of embryonic and foetal death.

Risk of clinical mastitis in dairy herds with a high proportion of low individual milk somatic-cell counts.

The relationships between the herd-somatic-cell count (SCC) pattern on a test day and the incidence of clinical mastitis in the subsequent period were studied by collecting health data and monthly records from 2 years (1995-1997) from 121 farms located in the west of France. A total of 980 herd-periods (from one test day to the following one) were analyzed. The outcome variable was the incidence density of clinical mastitis (ICM) within the herd-period. The herd-SCC pattern on a test day was described by cross-combining the proportions of cows with low SCC (<50,000 cells/ml) or with high SCC (>250,000 cells/ml). The relationship between herd-SCC pattern and subsequent ICM was assessed using a generalized linear mixed model. A sensitivity analysis evaluated the effect of different proportions of cows with low SCC on significance and magnitude of the relationship. Risk of clinical mastitis was expressed as a risk ratio (RR) in comparison to a moderate herd-SCC level (with low proportions of cows with low or high SCC). Median ICM was 0.38 cases per 365 cow-days at risk (first and third quartiles: 0 and 0.88). In the situations where few cows (<15%) had SCC>250,000 cells/ml, ICM was higher (RR>1.31) when the proportion of cows with low SCC exceeded 50% than in the reference situation. Risk ratios increased as the proportion of cows with low SCC increased from 40 to 60%. In the situations where the proportion of cows with SCC>250,000 cells/ml exceeded 15%, ICM was higher compared to the reference situation (whatever the proportion of low SCC). Risk ratios were also higher when the proportion of cows with low SCC were >40-60%, compared to the pattern with the proportion of cows with low SCC below 40%. Herd situations with a high proportion of cows with low SCC appeared to be at increased risk of clinical mastitis.