Benchmarks for evaluation and comparison of udder health status using monthly individual somatic cell count

Abstract

The objectives of this study were to propose benchmarks for the interpretation of herd udder health using monthly individual somatic cell counts (SCC) from dairy herds in Quebec, Canada and to evaluate the association of risk factors with intramammary infection (IMI) dynamics relative to these benchmarks. The mean and percentiles of indices related to udder infection status [e.g., proportion of healthy or chronically infected cows, cows cured and new IMI (NIMI) rate] during lactation and over the dry period were calculated using a threshold of ≥ 200 000 cells/mL at test day. Mean NIMI proportion and proportion of cows cured during lactation were 0.11 and 0.27. Benchmarks of 0.70 and 0.03 for healthy and chronically infected cows over the dry period were proposed. Season and herd mean SCC were risk factors influencing IMI dynamics during lactation and over the dry period.

Introduction

Improvement of milk quality and herd udder health are important for the dairy industry. There are high quality standards related to the milk produced on farms in Canada. From 1995 to August 2012, these standards specified that the bulk milk somatic cell count (BMSCC) had to be lower than 500 000 cells/mL. In August 2012, a quality program called “Canadian quality milk” was put in place for all Canadian farms and the upper limit for BMSCC was set at 400 000 cells/mL (1). Failure to comply with this standard would result in financial penalties and ultimately end milk shipment to the milk plant (2).

Among Quebec’s dairy herds in 2008, 14% of BMSCC were over 400 000 cells/mL and 5.7% were over 500 000 cells/mL. In 2011, 13% of BMSCC were over 400 000 cells/mL and 2.6% were over 500 000 cells/mL (3). An on-farm somatic cell count (SCC) monitoring program would be helpful for producers to monitor herd performance levels that comply with the new standards or to improve the udder health status of their herd.

The most important factor affecting SCC in milk of an individual quarter is the quarter infection status (4). Herd infection status and milk quality are related. The BMSCC is a good tool for fast and easy evaluation of milk quality but it is of limited use for evaluating infection dynamics within herds. For example, a high producing cow having a high SCC could cause a marked increase in the BMSCC of a relatively small herd. Looking only at BMSCC may give a false impression of generalized poor udder health within the herd (4). Monthly individual SCC could provide more precise information on herd mammary quarter infection dynamics for veterinarians, milk quality advisors, and producers evaluating herd udder health status. Although BMSCC is frequently used on the farm to evaluate udder health status, Lievaart et al (5) reported a higher coefficient of determination (R²) between the arithmetic mean of all the individual monthly SCC and the proportion of infected cows (cows with SCC > 250 000 cells/mL) in a herd (R² = 0.89) than between BMSCC and the proportion of infected cows (R² = 0.64). The yield-corrected test-day SCC and the proportion of infected cows had R² = 0.78. Madouasse et al (6) suggested that evaluation of monthly individual SCC and its transitions from one month to the next month can place a herd’s performance in context, identify problem areas, set targets, and assess the impact of management changes.

Few studies have explored interpreting monthly SCC to evaluate herd udder health status. In most of these studies, a threshold of 200 000 cells/mL was used to define intramammary infection (IMI) for estimating udder health status and setting target benchmarks (6–10). McDermott et al (11) and Dohoo and Leslie (12) suggested that this threshold had the best combination of specificity and sensitivity for defining an IMI.

Interpreting monthly SCC to evaluate herd udder health status was adopted by veterinarians in Quebec using DSAHR (Dossier Santé Animal — Animal Health Records) and CCStat program software (Zoetis, Kirkland, Quebec) in 2009. However, interpretation of the data is difficult because there are no established benchmarks for some of the monitored indices. A study was initiated in 2010 with the aim of identifying benchmarks for monthly individual SCC specific to dairy herds in Quebec, in order to facilitate interpretation of udder health status. The first objective was to estimate the proportion of IMI within herds in Quebec. The second objective was to estimate benchmarks related to the dynamics of healthy, cured, newly infected and chronically infected cows during lactation and over the dry period and for heifers infected at first test after calving. The third objective was to estimate the association of herd mean annual SCC, number of cows, herd mean daily milk yield per cow and season with the proportion of healthy, cured, newly infected and chronically infected cows in lactation and over the dry period and on the proportion of heifers infected at first test after calving.

Materials and methods

Database

A database including all DHI data from January 1 to December 31, 2008 of 950 dairy herds in Quebec was used. The entire database initially included 1729 herds; the final database consisted of 950 herds selected according to the following inclusion criteria. Herds had to be followed at least monthly by a veterinarian using the DSAHR software. Only herds with a minimum average size of 30 cows on DHI test and a minimum of 6 DHI tests during 2008 were included. The database included mainly Holstein cows with some Jersey, Ayrshire, and Brown Swiss. The final database was used for estimating the prevalence of infection at one test. Two data subsets were created for estimating the dynamics of infection during lactation and over the dry period, respectively. The 10th or 90th percentiles were estimated as target benchmarks of the best performing herds. The statistical analyses of indices were produced using SAS v. 9.2 (Cary, North Carolina, USA.) and Microsoft Excel.

Herd summary data

Mean annual number of cows per herd at test was calculated as the arithmetic mean of the number of cows at each monthly test per herd. The mean somatic cell count was established in the same way. The arithmetic mean linear score (LS) of the herd was calculated after transforming individual SCC for each test into individual LS and averaging them. Linear score was transformed using formula:

$$\text{LS}={log}_2(\text{SCC}/100)+3$$

Where: SCC is measured by 1000 cells/mL (13).

Estimating proportion of intramammary infection within herd

To evaluate the within herd proportion of infection, the herd median annual mean proportion of cows exceeding 200 000 cells/mL was calculated for the 950 herds during year 2008. Mean 10, 25, 75, 90 and minimum, maximum percentiles were also calculated. The proportion of cows with SCC < 200 000 cells/mL, ≥ 200 000 cells/mL, and < 500 000 cells/mL, ≥ 500 000 cells/mL, < 1 000 000 cells/mL and ≥ 1 000 000 cells/mL were calculated. These thresholds were selected because they are easily found with DSAHR software and are frequently used by Quebec veterinarians, milk quality advisors, and producers to evaluate the within herd proportion of IMI.

Estimating IMI dynamics during lactation

The complete dataset containing 117 770 cows in 1729 herds, of which, 78 656 cows in 949 herds met the inclusion criteria and constituted the final dataset. For 2 tests to be considered as consecutive tests, the interval between the 2 tests had to be 50 days or less. Descriptive statistics were estimated for the following indices, calculated at the annual level:

• proportion of healthy cows during lactation: number of cows with a SCC < 200 000 cells/mL in both of 2 consecutive tests divided by the total number of possible 2-consecutive tests individual SCC pairs;

• proportion of cows cured during lactation: number of cows with a SCC ≥ 200 000 cells/mL on the first of 2 consecutive tests within a pair and SCC < 200 000 cells/mL at the second of 2 consecutive tests within the same pair divided by the total number of possible 2-consecutive tests individual SCC pairs where the first test is ≥ 200 000 cells/mL;

• proportion of NIMI: number of cows with SCC < 200 000 cells/mL on the first of 2 consecutive tests within a pair and SCC ≥ 200 000 cells/mL at the second of 2 consecutive tests within the same pair divided by the total number of consecutive tests pair possibilities where the first test is < 200 000 cells/mL;

• proportion of chronic IMI: number of cows with a SCC ≥ 200 000 cells/mL in both of 2 consecutive tests divided by the total number of possible 2-consecutive tests individual SCC pairs.

Intramammary infection dynamics indices in lactation were analyzed globally and by subgroups of mean annual herd SCC (< 200 000 cells/mL; ≥ 200 000 cells/mL and < 400 000 cells/mL; ≥ 400 000 cells/mL), mean annual number of cows on test (< 40, ≥ 40 and < 60, ≥ 60), mean herd annual daily milk yield per cow (< 27 kg, ≥ 27 kg and < 29 kg, ≥ 29 kg and < 31 kg, ≥ 31 kg) and by season (season of the second test within each pair of 2 consecutive tests used). The association of these subgroups with the indices listed above was estimated using linear mixed models. Every index’s proportion was log transformed to normalize the distribution. Fixed effects in the models included mean annual herd SCC, mean annual herd linear score, mean annual number of cows on test, mean annual herd milk yield, and season. Herd was included as a random effect.

This dataset was also used to estimate the proportion of heifer IMI at first test, defined as the number of heifers with SCC ≥ 200 000 cells/mL at first test after calving divided by the total number of heifers with an eligible SCC record and a calving date in 2008. A first test SCC record was available if measured > 4 and ≤ 50 days after calving. Data were analyzed globally and by subgroups as was done for the evaluation of infection dynamics during lactation.

Estimating IMI dynamics over the dry period

To be included in the data-subset of SCC evaluation over the dry-period, cows had to experience a dry period longer than 30 d, a period of less than 150 d between the last test of its previous lactation and dry-off date and at least 1 test between > 4 and ≤ 50 days-in-milk in the subsequent lactation. This subset initially contained 38 718 cows having a complete dry period. The final subset contained 33 850 cow dry-periods in 949 herds that complied with the selection criteria.

Descriptive statistics were estimated for the following indices:

• proportion of healthy cows over the dry period: number of cows with SCC < 200 000 cells/mL before dry-off and after calving divided by the total number of cows dried off;

• proportion of cows cured over the dry period: number of cows with SCC ≥ 200 000 cells/mL before dry-off and < 200 000 cells/mL after calving divided by the total number of cows with SCC ≥ 200 000 cells/mL before dry-off;

• proportion of fresh cow NIMI: number of cows with SCC < 200 000 cells/mL before dry-off but ≥ 200 000 cells/mL after calving divided by the total number of cows with SCC < 200 000 cells/mL before dry-off;

• proportion of IMI persisting over the dry period: number of cows with a SCC ≥ 200 000 cells/mL before dry-off and after calving divided by the total number of cows dried off.

As was done for evaluating IMI dynamics indices during lactation, IMI dynamics indices over the dry period were analyzed globally and by subgroups. A similar linear mixed model as the one performed for the indices during lactation was used to evaluate the associations of the subgroups with the indices related to the dry period.

Results

Herd summary data

The annual median number of cows on DHI test was 49 (Figure 1). The median herd annual mean SCC was 262 000 cells/mL (Figure 2). The median herds mean annual daily milk yield was 29.3 kg/cow per day (Figure 3).

Figure 1.

Annual herd mean number of milking cows on test (n = 949). Quartile 25: 36 cows, median: 49 cows, quartile 75: 63 cows.

Figure 2.

Annual herd mean individual somatic cell count (SCC × 1000 cells/mL) (n = 949). Quartile 25: 200 000 cells/mL, median: 262 000 cells/mL, quartile 75: 328 000 cells/mL.

Figure 3.

Annual herd mean daily milk yield per cow (n = 949). Quartile 25: 27.0 kg/cow per day, median: 29.3 kg/cow per day, quartile 75: 31.4 kg/cow per day.

Proportion of infection within herd

The mean annual proportion of infection within herd (proportion of SCC ≥ 200 000 cells/mL) was 0.26 [95% confidence interval (CI) = 0.26 to 0.27]. The annual proportions of cows with SCC measures < 200 000 cells/mL, ≥ 200 000 cells/mL and < 500 000 cells/mL, ≥ 500 000 cells/mL and < 1 000 000 cells/mL and ≥ 1 000 000 cells/mL are shown in Table 1.

Table 1.

Annual herd mean and distribution (percentiles 10, 25, 50, 75, 90, minimum and maximum) of proportions of cow SCC measures < 200 000 cells/mL, ≥ 200 000 and < 500 000 cells/mL, ≥ 500 000 and < 1 000 000 cells/mL and ≥ 1 000 000 cells/mL in Quebec dairy herds during year 2008 (n = 950)

	Mean	95% CI	10	25	50	75	90	Min	Max
Proportion of cow SCC < 200 000 cells/mL	0.74	0.73–0.74	0.84	0.80	0.74	0.68	0.62	0.44	0.95
Proportion of cow SCC ≥ 200 000 and < 500 000 cells/mL	0.14	0.14–0.14	0.09	0.11	0.14	0.16	0.19	0.03	0.33
Proportion of cow SCC ≥ 500 000 and < 1 000 000 cells/mL	0.06	0.06–0.07	0.03	0.05	0.06	0.08	0.10	0.01	0.18
Proportion of cow SCC ≥ 1 000 000 cells/mL	0.06	0.06–0.06	0.03	0.04	0.06	0.08	0.10	0.00	0.20

95% CI — 95% confidence interval, Min — minimum, Max — maximum.

Intramammary infection dynamic during lactation

Intramammary infection dynamics during lactation are presented in Table 2. Data stratified by mean annual herd SCC and by season are presented in Table 3. Herds with higher mean SCC show less healthy cows and more chronically infected cows than herds with lower mean annual SCC (P < 0.05). There was no significant difference between subgroups of herd mean number of cows and herd mean daily milk yield per cow. The proportion of healthy cows during lactation was significantly higher during winter than during spring, summer, and fall and higher in spring than during summer and fall (P < 0.05). While no variation was found in the proportion of cured cows between seasons, the proportion of newly infected cows appeared higher during summer than all other seasons and higher in fall than during winter and spring (P < 0.05). The proportion of chronically infected cows was higher during summer than during winter and spring and higher in fall than all other seasons (P < 0.05).

Table 2.

Annual herd mean and distribution (percentiles 10, 25, 50, 75, 90, minimum and maximum) of proportions of healthy, cured, newly infected (NIMI) and chronically infected cows during lactation (n = 949)

	Mean	95% CI	10	25	50	75	90	Min	Max
Proportion of healthy cows	0.67	0.66–0.68	0.54	0.60	0.68	0.74	0.80	0.24	0.92
Proportion of cows cured	0.27	0.26–0.28	0.17	0.21	0.26	0.32	0.38	0.07	1.00
Proportion of NIMI	0.11	0.11–0.12	0.07	0.08	0.11	0.14	0.16	0.03	0.49
Proportion of chronic infection	0.19	0.18–0.19	0.09	0.13	0.18	0.23	0.29	0.00	0.49

95% CI — 95% confidence interval, Min — minimum, Max — maximum.

Table 3.

Annual herd mean proportions and standard deviations of healthy, cured, newly infected and chronically infected cows on 2 consecutives tests during lactation stratified by herd mean annual somatic cell count (SCC) and season of the second test (n = 949)

	Healthya	Curedb	Newly infectedc	Chronically infectedd
Herd mean SCC × 1000 cells/mL (SD5)
< 200 (n = 231)	0.78 (0.05)e	0.33 (0.10)e	0.07 (0.02)e	0.11 (0.04)e
200–400 (n = 617)	0.65 (0.07)f	0.26 (0.08)f	0.12 (0.03)f	0.20 (0.06)f
≥ 400 (n = 101)	0.52 (0.07)g	0.21 (0.06)f	0.17 (0.05)e,f	0.30 (0.06)g
Season
Winter	0.68 (0.11)e	0.28 (0.14)e	0.11 (0.055)e	0.18 (0.08)e
Spring	0.69 (0.11)f	0.27 (0.13)e	0.10 (0.049)e	0.18 (0.08)e
Summer	0.66 (0.11)g	0.28 (0.13)e	0.13 (0.058)f	0.19 (0.09)f
Fall	0.65 (0.11)g	0.27 (0.12)e	0.12 (0.057)g	0.20 (0.09)g

Means with different superscript letters (e–g) differ (P < 0.05).

^aProportion of healthy cows during lactation: number of cows with a SCC < 200 000 cells/mL in both of 2 consecutive tests divided by the total number of possible 2-consecutive test individual SCC pairs.

^bProportion of cows cured during lactation: number of cows with a SCC ≥ 200 000 cells/mL on the first of 2 consecutive tests within a pair and SCC < 200 000 cells/mL at the second of 2 consecutive test within the same pair divided by the total number of possible 2-consecutive test individual SCC pairs where the first test is ≥ 200 000 cells/mL.

^cProportion of NIMI: number of cows with SCC < 200 000 cells/mL on the first of 2 consecutive tests within a pair and SCC ≥ 200 000 cells/mL at the second of 2 consecutive tests within the same pair divided by the total number of consecutive tests pair possibilities where the first test is < 200 000 cells/mL.

^dProportion of chronic IMI: number of cows with a SCC ≥ 200 000 cells/mL in both of 2 consecutive tests divided by the total number of possible 2-consecutive test individual SCC pairs.

^eStandard deviation.

Intramammary infection dynamics over the dry period

Intramammary infection dynamics over the dry period is presented in Table 4. Data stratified by mean annual herd SCC and by season are presented in Table 5. Herds having a mean annual SCC < 200 000 cells/mL showed more healthy and cured cows and less chronically infected cows over the dry period than herds having a SCC ≥ 200 000 cells/mL (P < 0.05). No significant difference was found in the proportion of newly infected cows over the dry period between subgroups of herd mean annual SCC. No significant variation between subgroups of herd mean number of cows and herd mean daily milk yield per cow was found. Season showed no association with the proportion of healthy and chronically infected cows over the dry period. The proportion of cured cows over the dry period was significantly lower during summer than during winter and spring while it was not different from fall (P < 0.05). The proportion of newly infected cows over the dry period was significantly lower during winter than during summer and fall while it was not different from spring (P < 0.05).

Table 4.

Annual herd mean and distribution (percentiles 10, 25, 50, 75, 90, minimum, and maximum) of proportions of healthy, cured, newly infected and chronically infected cows over the dry period and annual proportion of heifers infected at first test after calving (n = 949)

	Mean	95% CI	10	25	50	75	90	Min	Max
Proportion of healthy cows	0.52	0.51–0.53	0.33	0.42	0.53	0.62	0.70	0.00	0.96
Proportion of cows cured	0.65	0.63–0.66	0.40	0.52	0.67	0.77	0.88	0.00	1.00
Proportion of new intramammary infection	0.20	0.19–0.21	0.07	0.12	0.18	0.26	0.35	0.00	0.80
Proportion of chronic infection	0.13	0.12–0.14	0.03	0.07	0.12	0.18	0.24	0.00	0.55
Proportion of heifers infection at calving	0.17	0.16–0.18	0.00	0.08	0.16	0.25	0.33	0.00	0.80

95% CI — 95% confidence interval, Min — minimum, Max — maximum.

Table 5.

Annual herd mean proportions and standard deviations of healthy, cured, newly infected and chronically infected cows on 2 consecutive tests over the dry period stratified by herd mean annual somatic cell count (SCC) and calving season (n = 949)

	Healthya	Curedb	Newly infectedc	Chronically infectedd
Herd mean SCC × 1000 cells/mL (SD5)
< 200 (n = 231)	0.65 (0.114)e	0.74 (0.19)a	0.15 (0.09)e	0.06 (0.05)e
200–400 (n = 617)	0.50 (0.121)f	0.63 (0.17)f	0.20 (0.11)e	0.14 (0.07)f
≥ 400 (n = 101)	0.37 (0.116)e,f	0.53 (0.17)f	0.27 (0.16)e	0.27 (0.10)f
Season
Winter	0.53 (0.24)e	0.67 (0.33)e	0.17 (0.22)e	0.12 (0.15)e
Spring	0.51 (0.24)e	0.65 (0.33)e	0.19 (0.22)e,f	0.13 (0.16)e
Summer	0.52 (0.23)e	0.60 (0.33)f	0.21 (0.21)f	0.14 (0.15)e
Fall	0.52 (0.22)e	0.64 (0.32)e,f	0.20 (0.21)f	0.13 (0.15)e

Means with different superscript letters (e,f) differ (P < 0.05).

^aProportion of healthy cows over the dry period: number of cows with SCC < 200 000 cells/mL before dry-off and after calving divided by the total number of cows dried off.

^bProportion of cows cured over the dry period: number of cows with SCC ≥ 200 000 cells/mL before dry-off and SCC < 200 000 cells/mL after calving divided by the total number of cows with SCC ≥ 200 000 cells/mL before dry-off.

^cProportion of fresh cow NIMI: number of cows with SCC < 200 000 before dry-off but SCC ≥ 200 000 cells/mL after calving divided by the total number of cows with SCC < 200 000 cells/mL before dry-off.

^dProportion of IMI persisting over the dry period: number of cows with a SCC ≥ 200 000 cells/mL before dry-off and after calving divided by the total number of cows dried off.

^eStandard deviation.

The proportion of heifer IMI at calving was 0.16 with a minimum of 0 and a maximum of 0.80 (Table 4). Data stratified by calving season are presented in Table 6. No significant difference was found in the proportion of infected heifers between subgroups of herd mean annual SCC, herd mean annual number of cows and herd mean annual daily milk yield per cow. There was a significantly higher proportion of infected heifers at calving in summer than in spring and fall while it was not different from winter. Winter showed a significantly greater proportion of these infections than fall (P < 0.05).

Table 6.

Annual herd mean proportions and standard deviations of infected heifers in early lactation stratified by calving season (n = 943)

Season	Infected heifers at calvinga
Winter	0.18 (0.28)b,c
Spring	0.16 (0.22)b,d
Summer	0.19 (0.24)c
Fall	0.15 (0.20)d

Means with different superscript letters (b–d) differ (P < 0.05).

^aNumber of heifers with SCC ≥ 200 000 cells/mL at first test after calving divided by the total number of heifers with an eligible SCC record.

Discussion

Even if the selection criteria (minimum of 30 cows and 6 DHI tests) may lead to a selection bias, the 950 herds of the database are representative of the 6601 dairy herds in Quebec in 2008. For all Quebec herds in 2008, the mean annual SCC was 257 438 cells/mL while the mean SCC of the database was 262 000 cells/mL (1). The Canadian Dairy Commission (CDC) estimated the mean number of cows per herd in 2008, at 55 cows for the year 2008 while the database herds showed a mean of 49 cows (14). The CDC estimation included all the cows present on farm while the database’s mean only included the cows tested at DHI.

Results presented in Tables 1 to 5 are useful to evaluate herds’ udder health performance. These tables allow comparison of indices from a specific herd to percentiles and benchmarks from a large databank. Strategies to improve udder health could be recommended and implemented. Managers in herds with values around the median should remain vigilant and try to put in place some improvement strategies. On the other hand, managers in herds with many values under the median value should spend more effort on characterization of the problem and the underlying causes and improving the situation.

Many similarities and some differences were found between the values and benchmarks presented in this study and published results in the literature. Table 7 presents a comparison between the study results and these benchmarks. Note that in the literature, some authors reported mean values while others presented median values. The authors of the present study chose to include both to facilitate the comparison and give the reader the option to use one or the other.

Table 7.

Quebec benchmarks summary and comparison with literature benchmarks

Percentage of…	Quebec (%)	Literature (%)
Intramammary infection	15	10b/20a
Healthy cows during lactation	80	NA
Cows cured during lactation	38	50a,e
New intramammary infection during lactation	7	10a/5b
Chronic infection during lactation	9	5b
Healthy cows over the dry period	70	NA
Cows curing over the dry period	88	80a/90b
Fresh cow new infections	7	5b
Infection persisting over the dry period	3	8c/3.3d
Heifer infections at first test	7	10a

^aSchukken et al (8).

^bBradley and Green (9).

^cSchukken et al (10).

^dMadouasse et al (6).

^eProportion of cows cured during lactation if treated.

The proportion of NIMI between two consecutive test-days during lactation was observed monthly by Cook et al (7). The authors reported NIMI ranging from 12.9% to 18.9%. Quebec’s dairy herds had a slightly lower mean proportion of NIMI during lactation (11%). The higher proportion of NIMI found in the study by Cook et al (7) may be explained by the fact that the proportion of NIMI was defined as the proportion of new infection (SCC > 200 000 cells/mL) at the current test in cows that were uninfected (SCC < 200 000 cells/mL) or non-lactating at the previous test while the definition used in this study did not include the non-lactating cows at previous test. Madouasse et al (6) calculated 13.8% NIMI during lactation for multiparous cows and 7.3% NIMI for primiparous cows (5). This classification by lactation number was not done in the present study.

The study herein had a lower proportion of cows cured during lactation compared with the proportion in the study by Madouasse et al (15) in which it was 35.8%. The percentages of cows chronically infected during lactation were estimated by Madouasse et al (6) to be 5.3% and 17.6% of chronically infected primiparous and multiparous cows, respectively. Quebec’s dairy herds had fewer cures and consequently more chronically infected cows than did dairy herds in England and Wales. This could partly be explained by different management practices in UK compared with Quebec. Motivation of UK farmers to control BMSCC may be higher since an economic penalty is applied on milk shipping when BMSCC is above 200 000 cells/mL (6) while penalties in Quebec in 2008 were only applied when the BMSCC exceeded 500 000 cells/mL. Note that the cure rate definition in this study and in the studies conducted by Madouasse et al (6,15) were based on the same DHI test intervals therefore allowing comparison between studies.

This study is the first to report the proportion of healthy cows over the dry period. A high proportion of healthy cows during the dry period could be an indicator of immune response efficacy and of a well-managed environment around calving. Housing, cleanliness, stress, and comfort management during the dry period are important to minimize teat contamination and further infection.

Madouasse et al (6) reported a higher percentage of cures (71.4%) during the dry period. Again, this difference may be partly due to the financial penalties applied to UK farmers exceeding a BMSCC of 200 000 cells/mL. The mean proportion of cows apparently cured over the dry period was estimated at 62.9% by Cook et al (7) with a wide range from 20% to 100%. A wide range was also found in the present study. Farm management factors (intramammary dry-cow antibiotic treatment, housing) probably contribute to this large variation between herds. In a Belgian study that included 14 766 heifers, De Vliegher et al (16) observed that 27.5% of heifers had IMI from day 4 to day 15 after calving. The same index was estimated by Madouasse et al (6) at 18.9% of heifers calving with a SCC > 200 000 cells/mL with a measurement period from 0 to 29 days in milk (6). This proportion was similar to what was found in the present study. The higher proportion found by DeVligher et al (16) is possibly due to a shorter risk period after calving than in the present study (> 4 and ≤ 50 days in milk). Schepers et al (17) reported that the SCC is usually elevated in the first days after calving and decreases to its minimal value between 40 and 80 days after calving.

As was expected, during lactation the proportion of healthy and cured cows significantly decreased and the proportion of newly and chronically infected cows significantly increased with increasing mean herd annual SCC. The most important factor affecting the SCC of the milk from an individual quarter, and consequently the cow and the herd, is the infection status of the quarter (4). Consequently, herds with more infection had higher mean SCC. The same pattern was found for the infection dynamics over the dry period.

It is known that SCC increases during summer months and decreases during winter months (4). Higher SCC is probably caused by increased contamination of the udder during months of elevated temperature and humidity and by decreased time devoted to cows by producers (4). A similar pattern is observed in this study with proportion of healthy cows during lactation being significantly lower in summer and fall than in winter and spring, and the proportion of newly infected and chronically infected cows during lactation being significantly higher during summer and fall than during winter and spring. The proportion of cured cows did not vary with the season and this could be explained by the well-established habit of milk producers to treat infected cows, regardless of the season. The association with season was less important in infection dynamics over the dry period. The proportion of cured cows over the dry period appeared to be significantly lower in summer than in winter and spring and the opposite pattern was seen with the proportion of newly infected cows over the dry period. Note that only the season of the first test after calving was included in the analysis. As the mean dry period length is about 60 d, it is possible for the dry period to range from one season to the next. For this reason, the interpretation of the results concerning the association between season and infection dynamic over the dry period is difficult.

The higher proportion of infected heifers at calving that was found during summer supports findings by Oliver et al (18) that summer season and absence of fly control are risk factors for heifer infection. However, the proportion of heifer infection at calving was not different between winter and summer. Fox et al (19) reported that winter months were associated with more heifer infection than other months of the year. Higher infection during winter could be caused by indoor housing management of heifers during the period before calving leading to higher risks of udder infection. Both summer and winter risk factors seem to affect the proportion of infected heifers in Quebec dairy herds.

In conclusion, this study provides benchmarks that can be used to evaluate herd udder health status based on the interpretation of monthly individual SCC in Quebec herds or comparable dairy herd populations. These benchmarks are useful to place a herd’s performance in context, identify problem areas, and set target actions to help producers achieve their goals concerning udder health.