Effectiveness of action in India to reduce exposure of Gyps vultures to the toxic veterinary drug diclofenac

Abstract

Contamination of their carrion food supply with the non-steroidal anti-inflammatory drug diclofenac has caused rapid population declines across the Indian subcontinent of three species of Gyps vultures endemic to South Asia. The governments of India, Pakistan and Nepal took action in 2006 to prevent the veterinary use of diclofenac on domesticated livestock, the route by which contamination occurs. We analyse data from three surveys of the prevalence and concentration of diclofenac residues in carcasses of domesticated ungulates in India, carried out before and after the implementation of a ban on veterinary use. There was little change in the prevalence and concentration of diclofenac between a survey before the ban and one conducted soon after its implementation, with the percentage of carcasses containing diclofenac in these surveys estimated at 10.8 and 10.7%, respectively. However, both the prevalence and concentration of diclofenac had fallen markedly 7-31 months after the implementation of the ban, with the true prevalence in this third survey estimated at 6.5%. Modelling of the impact of this reduction in diclofenac on the expected rate of decline of the oriental white-backed vulture (Gyps bengalensis) in India indicates that the decline rate has decreased to 40% of the rate before the ban, but is still likely to be rapid (about 18% year(-1)). Hence, further efforts to remove diclofenac from vulture food are still needed if the future recovery or successful reintroduction of vultures is to be feasible.

Figure 1. Locations of sampling site clusters in India.

The map shows centroids of 21 site clusters at which liver samples were obtained from carcasses of domesticated ungulates. Numbers next to the symbols identify site clusters listed in Table 1. Triangles show clusters sampled in all three surveys (T1, T2, T3), squares show clusters sampled in T1 and T2, diamonds, T1 and T3, and circles T1 only.

Figure 2. Comparison of the distributions of diclofenac concentrations before and after the ban on the veterinary use of diclofenac.

Cumulative distributions of diclofenac concentration (ppm wet weight) in ungulate liver samples from three surveys: red = T1, pre-ban, green = T2, soon after the ban, blue = T3, 7–31 months after the ban are shown by the stepped lines. The curves show cumulative Weibull distributions fitted separately to the data for each survey. Fitted values of prevalence f, the scale a and shape b parameters respectively were T1, 0.110, 1.336 and 0.592; T2, 0.122, 1.458 and 0.597; T3, 0.061, 1.844 and 0.673.

Figure 3. Comparison of probability density functions of diclofenac concentrations in ungulate liver before and after the ban on the veterinary use of diclofenac.

Fitted probability density functions are shown of diclofenac concentration (ppm wet weight) in ungulate liver samples from three surveys: red = T1, pre-ban, dark green = T2, soon after the ban, dark blue = T3, 7–31 months after the ban. The curves are derived from a Weibull model in which both the true prevalence of diclofenac f (including those with concentrations < LOQ) and the scale parameter a of the Weibull distribution of concentrations of diclofenac in those samples are determined by a site-cluster effect and a survey period effect. The shape parameter b of the Weibull distribution is assumed not to vary with site-cluster or survey period. Values of f and a in all three surveys were adjusted so that the results simulate those expected if the 21 site-clusters covered by the T1 (pre-ban) survey had been covered at the same sampling intensity in the second T2 and third T3 surveys.

Figure 4. Comparison of probability density functions of diclofenac dose per unit vulture body weight from ungulate tissue before and after the ban on the veterinary use of diclofenac.

Probability density functions are shown of estimated diclofenac dose (mg kg⁻¹ wet weight) per meal for birds eating a mixture of all edible ungulate tissues and feeding at intervals of three days. Results are shown for three surveys: red = T1, pre-ban, dark green = T2, soon after the ban, dark blue = T3, 7–31 months after the ban. The proportion of vultures expected to be killed by a given dose of diclofenac is shown by the dose-response curve (black, with right-hand y axis). The products of the dose probability density functions and the dose-response curve are shown by the orange, light green and light blue curves for surveys T1, T2 and T3 respectively. The areas under these curves give the estimated proportion of vultures killed per meal.

Figure 5. Changes in the expected rate of decline of the oriental white-backed vulture population in India.

Circles show the estimated rate of population decline, as a ratio relative to that determined from the T1 survey results in 2004 – 2005. Values are plotted at the mean sampling time for each of the surveys. Horizontal rectangles show the duration of the period covered by the sample collection for each survey. Vertical lines show 95% confidence limits for the ratios. Each of the four adjacent points in a set represents the result for a combination of assumptions (from left to right: F = 2, S₀ = 0.90; F = 2, S₀ = 0.97; F = 3, S₀ = 0.90; F = 3, S₀ = 0.97). Arrows show the timing of the recommendation by the National Board for Wildlife for a ban on the veterinary use of diclofenac and the withdrawal by the Drug Controller General of manufacturing licences for veterinary formulations of the drug.