Rate of decline of the Oriental white-backed vulture population in India estimated from a survey of diclofenac residues in carcasses of ungulates

Abstract

The non-steroidal anti-inflammatory drug diclofenac is a major cause of the rapid declines in the Indian subcontinent of three species of vultures endemic to South Asia. The drug causes kidney failure and death in vultures. Exposure probably arises through vultures feeding on carcasses of domesticated ungulates treated with the drug. However, before the study reported here, it had not been established from field surveys of ungulate carcasses that a sufficient proportion was contaminated to cause the observed declines. We surveyed diclofenac concentrations in samples of liver from carcasses of domesticated ungulates in India in 2004-2005. We estimated the concentration of diclofenac in tissues available to vultures, relative to that in liver, and the proportion of vultures killed after feeding on a carcass with a known level of contamination. We assessed the impact of this mortality on vulture population trend with a population model. We expected levels of diclofenac found in ungulate carcasses in 2004-2005 to cause oriental white-backed vulture population declines of 80-99% per year, depending upon the assumptions used in the model. This compares with an observed rate of decline, from road transect counts, of 48% per year in 2000-2003. The precision of the estimate based upon carcass surveys is low and the two types of estimate were not significantly different. Our analyses indicate that the level of diclofenac contamination found in carcasses of domesticated ungulates in 2004-2005 was sufficient to account for the observed rapid decline of the oriental white-backed vulture in India. The methods we describe could be used again to assess changes in the effect on vulture population trend of diclofenac and similar drugs. In this way, the effectiveness of the recent ban in India on the manufacture and importation of diclofenac for veterinary use could be monitored.

Figure 1. Locations of sites studied in India.

Sites from which liver samples were obtained from carcasses of domesticated ungulates in 2004–2005 for diclofenac assays are shown by circles (n = 67) and centroids of 73 road transect surveys used to measure the population trend of the oriental white-backed vulture in 2000–2003 are shown by stars.

Figure 2. Cumulative distribution of diclofenac concentrations in liver samples.

The stepped line shows the observed cumulative distribution of concentrations for 1,848 liver samples. Also shown is the fitted cumulative log-normal distribution in which the mean of log_e-transformed values is −1.1522 and standard deviation is 1.7670 (thick curve). The thin curve is the fitted third order complementary log-log model in which the cumulative probability is 0.8765+0.1235 (1–exp(-exp(0.3184+0.5415 log_e(d_liver)+0.05110 (log_e (d_liver))²+0.005058 (log_e(d_liver))³)).

Figure 3. Relationship of diclofenac concentration in selected tissues to that in liver from the same ungulate.

Log-log plots are shown of the concentration of diclofenac in intestine, fat, kidney and muscle against that in the liver. Symbols denote ungulate species and data source: open diamonds-Bos indicus, Experiment 1 ; circles-Bos indicus, Indian carcass dumps; squares-Bos taurus, Experiment 2; triangles-Bos taurus, Experiment 3; grey diamond–Bubalus bubalis, . Lines show results from the fitted Model E, in which the geometric mean concentration in the selected tissue is assumed to be a fixed proportion of the concentration in liver (k = 1, see text).

Figure 4. Relationship between the proportion of oriental white-backed vultures treated experimentally with diclofenac that died and the dose of diclofenac administered per unit vulture body weight.

Plotted points are proportions killed for each of five bins of dose, with 95% exact binomial confidence limits (vertical lines). Bins include 6, 4, 2, 10, and 2 birds respectively (ranked from lowest to highest dose). The thick curve is the fitted probit model relating mortality rate to log dose. Thin curves show the envelope enclosing the central 9,500 of 10,000 Monte Carlo replicate values. The upper panel shows results of analysis of all data. The lower panel excludes the datum for an outlier (vulture Gb11) from the lowest dose bin, which died even though it apparently received an extremely low dose of diclofenac. Data are from Table 2 of Oaks et al. .