The use of common swift (Apus apus), an aerial feeder bird, as a bioindicator of persistent organic microcontaminants

Abstract

In this study we investigated the accumulation of polychlorobiphenyls (PCBs), polychlorodibenzo-p-dioxins (PCDDs), polychlorodibenzofurans (PCDFs), and the chlorinated pesticides 1,1-dichloro-2,2-bis(4-chlorophenyl)-ethene (DDE), 1,1,1-trichloro-2,2-bis(4-chlorophenyl)-ethane (DDT), and hexachlorobenzene (HCB) in the breast muscle, liver, lung, heart and brain tissues of adult common swifts (Apus apus, a long-lived aerial feeder bird). Individuals were collected in an urban area located in Rome during the breeding period. As shown by lipid-base normalized data, in general analytes had a significant minimum concentration in the brain. PCDD and PCDF concentration values in such tissue were approximately one order of magnitude lower than those found in breast muscle, heart, and lung tissues, and as much as two orders of magnitude below the relatively high levels found in the liver. PCB levels followed the same accumulation patterns. Of all analytes, HCB exhibited the most uniform distribution pattern over the five matrices assayed. DDE and DDT were by far the most and the least concentrated pesticide. In the urban environment of Rome, an air-to-swift bioconcentration factor (lipid based) in the order of 5 x 10(6) (2 x 10(5), fresh tissue base) was estimated for PCDDs and PCDFs. Our study suggests that airborne arthropod feeders, such us the common swift, are suitable biomonitors for air quality assessment.

Full text

187
Ann Ist super sAnItà 2008 | Vol. 44, no. 2: 187-194
reseArch from AnImAl testIng to clInIcAl experIence
Address for correspondence: Alessandro di Domenico, Dipartimento di Ambiente e Connessa Prevenzione Primaria, Istituto
Superiore di Sanità, Viale Regina Elena 299, 00161 Rome, Italy. E-mail: addeke@iss.it.
Summary. In this study we investigated the accumulation of polychlorobiphenyls (PCBs), polychlo-
rodibenzo-p-dioxins (PCDDs), polychlorodibenzofurans (PCDFs), and the chlorinated pesticides
1,1-dichloro-2,2-bis(4-chlorophenyl)-ethene (DDE), 1,1,1-trichloro-2,2-bis(4-chlorophenyl)-ethane
(DDT), and hexachlorobenzene (HCB) in the breast muscle, liver, lung, heart and brain tissues of
adult common swifts (Apus apus, a long-lived aerial feeder bird). Individuals were collected in an
urban area located in Rome during the breeding period. As shown by lipid-base normalized data, in
general analytes had a signicant minimum concentration in the brain. PCDD and PCDF concen-
tration values in such tissue were approximately one order of magnitude lower than those found in
breast muscle, heart, and lung tissues, and as much as two orders of magnitude below the relatively
high levels found in the liver. PCB levels followed the same accumulation patterns. Of all analytes,
HCB exhibited the most uniform distribution pattern over the ve matrices assayed. DDE and DDT
were by far the most and the least concentrated pesticide. In the urban environment of Rome, an
air-to-swift bioconcentration factor (lipid based) in the order of 5 × 106 (2 × 105, fresh tissue base)
was estimated for PCDDs and PCDFs. Our study suggests that airborne arthropod feeders, such us
the common swift, are suitable biomonitors for air quality assessment.
Key words: organochlorine contaminants, POPs, air contamination, bioindicators, common swift, PCDDs and PCDFs.
Riassunto (Il rondone comune “Apus apus, Aves”, un insettivoro aereo specializzato, come biondi-
catore di contaminanti organici persistenti). È stata quanticata la presenza nel rondone comune
(Apus apus) di policlorobifenili (PCBs), policlodibenzodiossine (PCDDs), policlorodibenzofurani
(PCDFs), e dei pesticidi clorurati 1,1-dicloro-2,2-bis (4-clorofenil)-etene (DDE), 1,1,1-tricloro-2,2-
bis (4-clorofenil)-etano (DDT), e esaclorobenzene (HCB). Esemplari di rondoni adulti sono stati
raccolti a Roma durante il periodo riproduttivo. Come mostrato dai dati normalizzati sul grasso
corporeo, in generale gli analiti hanno una concentrazione minima signicativa nel cervello. Le con-
centrazioni di PCDD e PCDF in questo tessuto risultavano approssimativamente un ordine di gran-
dezza inferiori a quelle trovate in muscolo pettorale, cuore e tessuto polmonare, e no a due ordini di
grandezza inferiori ai livelli relativamente alti trovati nel fegato. I livelli di PCB misurati nel cervello
avevano le stesse caratteristiche di distribuzione. Di tutti gli analiti, l’HCB esibiva la distribuzione
più uniforme di contaminazione tra tutte le matrici analizzate mentre DDE) e DDT erano di gran
lunga i pesticidi più e meno concentrati. Nell’ambiente urbano di Roma è stato stimato un fattore di
bioconcentrazione aria-rondone per PCDDs e PCDFs dell’ordine di 5 × 106 (2 × 105, peso fresco).
Parole chiave: contaminanti organoclorurati, contaminanti organici persistenti, contaminazione dell’aria, bio-
indicatori, rondone comune, policlorodibenzo-p-diossine, policlorodibenzofurani.
INTRODUCTION
The ubiquitous polychlorobiphenyls (PCBs), poly-
chlorodibenzo-p-dioxins (PCDDs), polychlorodiben-
zofurans (PCDFs), and chlorinated pesticides such as
1,1-dichloro-2,2-bis (4-chlorophenyl)-ethene (DDE),
1,1,1-trichloro-2,2-bis(4-chlorophenyl)-ethane
(DDT), and hexachlorobenzene (HCB) are lipophilic
organic microcontaminants of the environment char-
acterized by a remarkable environmental persistence,
potential for bioconcentration, and a range of pos-
sible adverse effects on wildlife populations including
carcinogenicity and interference with the endocrine
system [1, 2]. Furthermore, in humans several of these
compounds exhibit a strong toxic action, including
carcinogenity, even at very low exposure levels [3-8].
Many authors have suggested the use of birds as
The use of common swift (Apus apus),
an aerial feeder bird, as a bioindicator
of persistent organic microcontaminants
Roberto Miniero(a), Claudio Carere(b), Elena De Felip(a), Nicola Iacovella(a),
Fabrizio Rodriguez(a), Enrico Alleva(b) and Alessandro di Domenico(a)
(a)Dipartimento di Ambiente e Connessa Prevenzione Primaria;
(b)Dipartimento di Biologia Cellulare e Neuroscienze, Istituto Superiore di Sanità, Rome, Italy

188 Roberto Miniero, Claudio Carere, Elena De Felip, et al.
monitors of pollutants in terrestrial and especially
aquatic environments, and sh-eating birds or rap-
tors high on the foodchain have been primarily used
[9]. Some studies have demonstrated that insectivo-
rous passerine birds are suitable monitors for air,
heavy metal, and organochlorine pollution [10, 11].
However, to our knowledge no effort has been at-
tempted to nd bioindicators specic for the urban
environment and air pollution, while the number of
species/populations adopting synanthropic habits
has markedly increased in the last decades [12, 13].
Among animals, birds probably represent the most
successfully adapted to synanthropic life, showing
a rapid radiation on co-adaptive use of human by-
products. Many species with different breeding and
feeding styles continuously use human settlements
and man-made landscape and share with human
populations a somehow similar pattern of exposure
to xenobiotics.
The long-lived (up to 12 years) common swift (Apus
apus, order Apodiformes) breeds in colonies, pref-
erentially in town and village buildings, and spends
most of its active time ying. It feeds exclusively on
airborne arthropods, and mostly in the range from
approximately ground level to 100 m above open
ground [14]. The species is a sexually monomorphic,
long-distance migratory bird (weight: 38-42 g), whose
western Palearctic populations move regularly from
sub-Saharan Africa over to Europe for breeding [15].
At our latitudes (Rome, Italy, 41° 44’ N, 12° 24’ E),
they arrive in early spring to spend approximately
four months (April-July) from colony establishment
to edging of the offspring. Very little is known of
those periods when swifts are presumably in central
and southern Africa [15]. The chief asset of the com-
mon swift as bioindicator is its peculiar aerial niche
occupied in the urban and peri-urban environment,
together with its abundance and wide distribution
permitting sampling in almost any urban area across
much of Europe.
The present study is a follow-up of a preliminary
investigation [16]. We aim: (a) to evaluate whether
the common swift has a potential for use as a bioindi-
cator of the chemicals of interest; (b) to assess how
exposure, body burden, and distribution in different
organs are related. Moreover, in the same population
monitored, patterns of parental behaviour deviating
from the expectation (bi-parental care vs maternal
care) have been reported [17], leading to hypothesize
the occurrence of endocrine disruption phenomena
[18-20]. The assessment of the above mentioned con-
taminants may serve as a rst step to indicate (or
not) causal links explaining the observed variation.
Results are described and discussed together with an
exposure model tentatively developed for the common
swift. As there is a substantial lack of relevant data
concerning airborne-feeding birds and their possible
exposure pattern(s) to the microcontaminants dealt
with here, this study, while serving to ll in informa-
tion gaps, is also intended to generate hypotheses and
provide suggestions for further research.
MATERIALS AND METHODS
Sample collection
Forty fatally injured grounded adult swifts were
collected in Rome between April 29 and July 14
1996, during the breeding period [16]. The injured
birds were transported to the laboratory and eutha-
nized with ethyl ether once it was ascertained they
had no chance of survival. The plumage and skin of
each body were carefully removed upon delivery to
the laboratory, and so were the brain, breast muscle,
heart, liver, and lungs. Instruments were carefully
cleaned with hexane after each specimen’s dissec-
tion. Individual specimens were wrapped in alumi-
num foil, identied, and stored at -80 °C awaiting
further treatments.
Analysis
Fifteen randomly-selected specimens from the
entire specimen batch available were allowed to
thaw out in the laboratory; then, ve tissue-specic
samples were obtained by pooling the specimens
of a type. Samples were combined with a 1:1 (v/v)
n-hexane-acetone mixture, spiked with 13C-labeled
standards (CIL, Cambridge Isotope Laboratories),
and homogenized and extracted by means of a me-
chanical homogenizer to remove the lipid compo-
nent. The crude organic extract was subjected to
a number of cleanup steps including liquid-liquid
partition, treatment with concentrated sulfuric ac-
id, and chromatographic ltration on an activated
alumina column. The two main eluates obtained
from alumina cleanup were used for determina-
tion of PCBs, DDE, DDT, and HCB (Fraction 1),
and PCDDs and PCDFs (Fraction 2) by HRGC-
HRMS(SIM). GLP and QA/QC protocols were
applied throughout; conrmatory determinations
were eventually carried out. Uncertainty of the GC-
MS measurements reported herein was estimated as
<CV> ≈ 10% (CV < 30%). The analytical procedure
was adopted from the USEPA Method No 1613
[21]. In the text the prexes T4, P5, H6, H7, and O8
have been used to indicate chlorosubstitution levels
of four through eight.
RESULTS AND DISCUSSION
Analytical outcomes
A synopsis of the relevant ndings obtained is pre-
sented in Table 1, where PCB and PCDD and PCDF
amounts are expressed as cumulative data. The fresh
tissue samples utilized for analytical assessments
weighed between 4.24 and 11.7 g. The lipid amounts
extracted ranged from 89.0 to 574 mg, between 2.10
and 5.82% of the original matrices. The lowest lipid
content was found in lung tissue, whereas the highest
was measured in the brain, as expected.
All the analytes appear to have a consistent
minimum concentration in the brain, as shown
specically by lipid-base normalized data (how-
ever, DDT concentration in the liver could not be
measured). In particular, PCDD and PCDF con-

189
BIrds As BIoIndIcAtors of persIstent orgAnIc pollutAnts
centration values in the brain (199 pg/g and 19.4
pg I-TE/g or 23.9 pg WHO-TE/g where I-TE and
WHO-TE indicate conversion of congener-specic
analytical data to 2,3,7,8-T4CDD toxicity equiva-
lents as per the I-TEF [22] and WHO-TEF [23]
systems are approximately one order of magnitude
lower than the pertinent levels found in breast
muscle, heart, and lung tissues, and up to two or-
ders of magnitude below the relatively high levels
found in the liver (27,900 pg/g and 763 pg I-TE/g
or 954 pg WHO-TE/g). In terms of concentration
drop magnitude, PCBs come next with an assessed
value in the brain of 1100 ng/g, some four to eight
times smaller than the concentration gures esti-
mated for the remaining tissues, all of which fall
in the comparatively narrow range of 4060-7560
ng/g. Aside from the aforecited unassessed DDT
level in the liver, pesticide tissue distributions seem
to be characterized by concentration levels not as
variable as those detected for PCBs and PCDDs
and PCDFs, with reduced differences between
pertinent maxima and minima whose ratios are
approximately within a factor of 4. Of the entire
group of analytes, HCB exhibits the most uniform
distribution pattern (93.6-229 ng/g), whereas DDE
and DDT are, respectively, by far the most and the
least concentrated pesticide in the ve tissues exam-
ined (specically, [DDT] << [HCB] << [DDE]).
When compared with the published literature, the
PCB levels in the upper range (Table 1) appear to be in
reasonable agreement with similar data obtained from
areas under general anthropogenic impact [24, 25], an
observation which also applies to the chlorinated pesti-
cides quantied.
Exposure analysis and model
Pesticide contamination levels and distribution pat-
terns may reect the fact that the organisms exposed
had the chance to reach or approach what will be here
referred to as a “steady-state-like condition”. This situ-
ation would apply particularly well to the exposure
period(s) of approximately seven months a year that
swifts spend in Africa, where DDT is still used [26].
However, exposure to DDT and its kin compounds and
metabolites – including the prominent DDE – is liable to
occur also in those countries (such as Italy) that banned
their use in the open long ago, but where the chemicals
are still environmentally present as historical residues,
not counting chemical accidents, other forms of local
releases, and the long-range transport. Indeed, it
Table 1 | Concentration levels a of the chlorinated microcontaminants assessed in ve selected tissue types of the common swift
(sampling from April 29 to July 14, 1996). Analyte values are expressed per unit fresh tissue weight and per unit extracted lipid weight
Analyte Analyte levels per tissue type
Brain Breast
muscle
Heart Liver Lungs
Data normalized on fresh tissue base (ng/g, except where noticed)
PCBsb64.2 371 155 225 100
PCDDs+PCDFsc (pg/g) 11.6 71.6 117 947 144
PCDDs+PCDFsc (pg I-TE/g) 1.13 8.27 6.92 25.8 4.27
PCDDs+PCDFsc (pg WHO-TE/g) 1.39 10.3 8.63 32.2 4.79
DDE 112 379 263 251 117
DDT 0.388 0.928 1.09 <0.034 0.377
HCB 5.45 8.21 6.65 7.76 4.02
Data normalized on lipid base (ng/g, except where noticed)
PCBsb1100 7560 4060 6680 4760
PCDDs+PCDFsc (pg/g) 199 1460 3060 27,900 6830
PCDDs+PCDFsc (pg I-TE/g) 19.4 168 181 763 203
PCDDs+PCDFsc (pg WHO-TE/g) 23.9 210 226 954 228
DDE 1930 7720 6900 7410 5560
DDT 6.67 18.9 28.7 <1.0 17.9
HCB 93.6 167 174 229 191
Ancillary information
Matrix size (g) 5.98 11.7 6.12 10.6 4.24
Extracted lipid (g) 0.348 0.574 0.234 0.357 0.0890
Extracted lipid (%) 5.82 4.91 3.82 3.37 2.10
(a)Values corrected for analytical recovery and rounded off to three gures. Individual gures of the original data sets preceded by ≤ or < (signs respec-
tively indicating an upper limit estimate or below limit of quantication, S/N ≈ 3; N ≈ 4 σn), were entered as half their nominal value to calculate PCB
and PCDD and PCDF cumulative data.
(b)As per PCB congeners reported in Table 2.
(c)As per PCDD and PCDF congeners reported in Table 3.

190 Roberto Miniero, Claudio Carere, Elena De Felip, et al.
may be presumed that, moving from Africa to Italy,
birds would be exposed to lesser amounts of the pes-
ticides of interest, this entailing a relatively slow re-
adjusting to a somewhat lower level of the previous-
ly reached steady-state-like condition. DDT absence
in the liver may be explained by the fact that most
of the substance metabolic fate is met there [3], the
remaining organs examined merely acting as storing
and/or supplying matrices. Toxic and stable DDE is
an important metabolite of DDT and an end prod-
uct itself in some metabolic systems [3, 4]: the DDE-
DDT pattern detected (Table 1) – for which a high
[DDE] × [DDT]–1 ratio is estimated (>>200) – could
indicate the occurrence of an old exposure [27], thus
supporting the hypothesis that birds were sampled
after a steady-state-like condition for the pesticides
was reached.
HCB is not only a widely used pesticide (fumi-
gant), but also a common intermediate or by-prod-
uct of important chloro-organic industrial produc-
tions [8]. Due to its comparatively greater volatility,
HCB is a worldwide atmospheric contaminant ex-
hibiting substantially uniform distribution levels.
Urban areas are important sources of PCDDs
and PCDFs, as proven by different studies around
the world [7] including those for the Italian cities of
Rome and Florence [28, 29]. Therefore, a signicant
uptake of the aforcited microcontaminants could be
associated with the comparatively short period(s)
the swifts are in Italy, thus determining a chemi-
cal distribution in the swifts collected that did not
have enough time to develop into the features of a
steady-state-like condition. The relatively high levels
of these substances found in the liver – a canonical
primary target organ [5] – together with the remark-
able unevenness of concentration levels in the tissues
examined may be thought of as supporting evidence
of such a hypothesis. In addition, it may be observed
that the long PCDD and PCDF half-lives (hls) in liv-
ing organisms – actually, the longest hls known for
xenobiotic chemicals – would favor a comparatively
rapid building-up of tissue levels (and, conversely,
a comparatively slow depletion), thereby stressing
inter-tissue concentration differences before the
steady-state-like condition is reached [7]. Finally, let
us assume that the swifts assayed were exposed to
PCDDs and PCDFs primarily in the urban environ-
ment of Rome. This assumption is based on the fact
that swifts show a high nest/colony delity and dur-
ing the whole breeding period adults have to return
to their nests every 2-4 h for the parental activities
[17, 30]. Then, the indicative PCDD and PCDF con-
centrations for the swifts (body burden) and the cor-
related media (ambient air) may respectively be tak-
en as 10 pg I-TE/g, fresh tissue base, derived from
a coarse weighting of tissue contributions (Table
1), and 5 × 10–5 pg I-TE/g (approximately, 0.07 pg
I-TE/m3) [28, 29]. The swift is a pure aerial feeder,
therefore an air-to-swift bioconcentration factor
(BCF) can be taken into consideration. The above-
indicated gures would yield an estimate of BCF in
the order of 2 × 105 or possibly greater (the air con-
centration values taken as a reference were obtained
mostly during the winter period, when home heating
was an additional source of contamination with re-
spect to the spring-summer period). If the lipid base
concentrations are considered (Table 1), the BCF
estimate is in the order of 5 × 106.
PCB distribution pattern seems to fall between the
two extreme cases described above: it is somewhat
more uneven than that of pesticides, but does not
show the variability of up to two orders of magni-
tude observed for PCDD and PCDF concentration
levels. Therefore, owing to the worldwide environ-
mental diffusion of PCBs, their distribution levels
and pattern in the common swift tissues analyzed
may reect substantially continuous exposures the
year-round – possibly, to some extent greater in Italy
[28, 29] than in Africa’s wintering areas – and may
also reect the presence in the organisms examined
of a steady-state-like condition nearly reached. In
particular, if moving from Africa to Italy would
have the birds exposed to greater environmental
PCB concentrations, a re-adjusting to a somewhat
higher level of the previously reached steady-state-
like condition, with a comparatively rapid build-
ing-up of PCB tissue levels, should be expected.
This process could stress inter-tissue concentration
differences while the organisms are shifting toward
a steady-state-like condition of a somehow higher
level. Similar to the pesticides, the industrially-pro-
duced PCBs are still environmentally present as his-
torical residues from a wide variety of commercial
uses that, for decades since the 1940s, determined
their largely uncontrolled release into the environ-
ment; today, their use is restricted to closed-cycle
applications (e.g., as dielectric uids in heavy-duty
electrical appliances) and their production is banned
almost anywhere [6]. Besides chemical accidents,
(improper) treatment and disposal of PCB-contain-
ing waste, and other forms of local releases however
uncommon, the widespread environmental presence
of the aforecited chemicals is also fueled by long-
range atmospheric transport.
Finally, it may be observed that the general analyte
concentration drop found in the brain (Table 1) is ac-
counted for by the well-known hindering action that
hematoencephalic (or blood-brain) barrier exerts on
xenobiotics or, in general, exogenous substances [31].
Analyte congener- and compound-specic proles
In Table 2, H6CB[138+163], H7CB[170], H7CB [180],
and H7CB[187] appear to supply rm contributions
larger than 20% of base congener H6CB[153] (rela-
tive abundance, 100%). This ve-congener cluster,
where H7CB[180] may be seen to compete as the
base congener, is a recurrent and remarkably stable
primary feature typical of all PCB distributions as-
sessed in the swift samples, and is typical of animals
sharing high trophic levels. PCB proles reported
in Table 2 and detected in the ve tissues examined
show a degree of high reciprocal similarity – a fact

191
BIrds As BIoIndIcAtors of persIstent orgAnIc pollutAnts
suggestive that a steady-state-like condition has
nearly come about. In particular, each value is as-
sociated with its coefcient (CV) and range of vari-
ation. Prole stability is borne out by the magnitude
of CV estimates, most of which (frequency, 15/29)
are less than 20%. In any case, the less chlorinated
homologs are known to have a metabolic reactivity
greater than those with a higher degree of chloro-
substitution [1]; therefore, the perceivable differenc-
es that characterize the occurrence in the different
tissues of several T3CB, T4CB, and P5CB congeners
could be tentatively ascribed to exposure or meta-
bolic factors, or a combination thereof.
As seen for PCBs, tissue-specic pesticide pro-
les were so similar to each other that they could
be made into a single data set of mean compound-
specic values. As stated before, this is additional
evidence that a steady-state-like condition may have
taken place. Again, prole stability is indicated by
the magnitude of DDE, DDT, and HCB CV esti-
mates, respectively 1.03, 21.1, and 31.2%. The com-
paratively large HCB CV value is at least in part at-
tributable to the chemical’s greater volatility, a fac-
tor that may inuence analytical repeatability.
From the tissue- and congener-specic data of
Table 3, it may readily be observed that PCDFs
consistently provide only a minor and quite variable
contribution (2.8-23%) to the overall cumulative an-
alytical amounts; their contribution to cumulative i-
te values is larger and not as spread (17-40%) but still
visibly smaller than the PCDD complement. From
the table, it may be seen that PCDD and PCDF
analytical proles are ngerprinted by the recurrent
presence of O8CDD – the base congener, account-
ing for some 30-70% of the pertinent cumulative
amounts – accompanied by 1,2,3,4,6,7,8-H7CDD
and 1,2,3,6,7,8-H6CDD, two congeners that eventu-
ally compete as the next most important constitu-
ent. The congener 2,3,4,7,8-P5CDF appears as the
most prominent of PCDFs, although in two cases
only (brain and breast muscle tissues) does it come
within the ve principal constituent group.
As stated previously, average PCB and pesticide
proles could be built by merging the original tis-
sue-specic ndings, owing to the good inter-tissue
similarity of individual proles. For PCDDs and
PCDFs, such an operation cannot be performed due
to their irregular distribution, at both congener and
cumulative (analytical) levels, in the ve tissues ex-
amined. This aspect is clearly borne out by the out-
comes (Table 4, “ve-tissue” part) of the analysis of
original congener data shown in Table 3. The lowest
CV values, 46.8 and 55.4%, are respectively associ-
ated with 2,3,7,8-T4CDD and 1,2,3,7,8,9-H6CDF. In
all the remaining cases, CVs are well above 60% and
often (frequency, 6/17) exceed 100%. The important
ngerprinting congeners O8CDD and 1,2,3,4,6,7,8-
H7CDD exhibit the largest CVs, 160 and 167%,
respectively. Anyway, the tissue-specic congener
accumulation seem to follow the pertinent increase
in partition coefcient values. As absolute analyti-
cal gures (pg/g, fresh tissue base) were used for this
evaluation, CV magnitude may be taken as an indi-
cator of inter-tissue distribution variability of indi-
vidual PCDDs and PCDFs (i.e., the larger the CV
value, the more irregular the distribution). On this
basis, it may be pointed out that the whole PCDF
congener group and, in particular, the PCDD and
Table 2 | PCB analytical prole descriptors obtained from original tissue- and congener-specic data sets available from the assay of
common swift brain, breast muscle, heart, liver, and lung specimens.a From the mean descriptor data set, a mean prole is obtained
Analyte Profile descriptorsbAnalyte Profile descriptorsb
Mean n cv% Min Max Mean n cv% Min Max
T4CB [66+80] 1.89 4 16.2 1.61 ≤10.9cH6CB [167] 4.67 4 9.88 ≤3.15 5.30
T4CB [74] 1.95 4 37.4 1.24 ≤12.7 H7CB [170] 26.7 4 6.88 ≤18.2 29.1
P5CB [87] 1.26 5 41.6 0.713 2.13 H7CB [171] 1.57 3 15.7 1.30 ≤2.05
P5CB [95] 2.52 4 55.6 ≤1.03 4.54 H7CB [174] 1.19 3 28.0 ≤0.759 1.57
P5CB [99] 3.60 4 34.0 ≤1.90 5.40 H7CB [177] 8.56 2 — ≤6.36 9.29
P5CB [101] 3.33 5 53.4 1.75 6.25 H7CB [178] 4.62 5 7.38 4.16 5.11
P5CB [105] 2.55 4 64.1 0.365 ≤5.77 H7CB [180] 92.2 5 14.1 70.9 106
P5CB [118] 18.3 5 20.9 16.2 25.1 H7CB [183] 9.41 4 7.80 ≤7.99 10.1
H6CB [138+163] 52.3 3 13.4 ≤35.7 59.5 H7CB [187] 32.9 5 8.68 29.3 37.0
H6CB [146] 10.9 3 9.12 ≤9.45 11.8 O8CB [194] 13.9 5 19.8 9.67 17.1
H6CB [149] 3.59 5 33.1 2.71 5.63 O8CB [195] 3.38 4 20.8 ≤2.36 4.27
H6CB [151] 1.16 5 30.2 0.839 1.74 O8CB [201] 11.9 5 15.6 9.54 14.1
H6CB [153] 100 4 5.89 ≤86.0 105 O8CB [202] 2.15 5 22.8 1.62 2.82
H6CB [156] 9.21 5 5.03 8.82 9.93 O8CB [203+196] 8.33 5 19.1 5.98 10.1
H6CB [157] 1.98 3 22.9 ≤0.791 2.37
(a) Sampling campaign, from April 29 to July 14, 1996.
(b)Mean, minimum, and maximum descriptor values are normalized on the base congener (H6CB [153]) mean estimate. N provides the number of origi-
nal entries utilized for congener data processing.
(c)The sign ≤ indicates an upper limit estimate. Data of this type were not entered in calculations, a fact that accounts for N < 5.

192 Roberto Miniero, Claudio Carere, Elena De Felip, et al.
PCDF congeners with lower chlorosubstitution,
seem to have a less irregular distribution than that
of, respectively, PCDDs and those congeners with a
higher number of chlorine. The ratios [max] × [min]–1
exhibit a pattern that is by and large in agreement
with that of CV estimates, an additional evidence
in support of the aforecited observations. Ratio val-
ues range from less than ve to over two orders of
magnitude (3.46-168), reecting the differences of
congener concentrations in liver and brain tissues,
consistently associated with respectively maximum
and minimum estimates. Exclusion from the data
analysis of concentration gures concerning the
brain (Table 4, “four-tissue” part) results in a sub-
stantial reduction of all CV estimates in spite of the
concurrent degree-of-freedom reduction (N, 5 × 4
or 4 × 3). The values of the ratio [max] × [min]–1
also appear signicantly decreased, more visibly
for PCDDs than for PCDFs. In either case, how-
ever, elimination of the lowest concentration g-
ures from the general data set still leaves congener
distribution patterns characterized by remarkable
amounts of spread (CV, from 31.5 to 146%; [max]
× [min]-1, 2.03-31.3). In other words, even if the ef-
fect of the hematoencephalic barrier is accounted
for, distributions of individual PCDDs and PCDFs
over the remaining four tissues are still very un-
even, more so for congeners with a higher number
of chlorine, a fact that may be reasonably explained
if exposure to these chemicals is comparatively re-
cent, and therefore a steady-state-like condition has
not been reached as yet. The observations reported
above hold for PCDD and PCDF cumulative levels
as well, as may be deduced from Table 4. Lastly, the
above discussion was based on PCDD and PCDF
outcomes normalized on fresh sample weights: the
use of lipid-base normalized data would have the
effect of increasing somewhat the inter-tissue dis-
tribution variability, leading however to evaluations
similar to those already described.
CONCLUSIVE REMARKS
In general terms, the use of the common swift as
a bioindicator is suggested especially to detect those
chemicals that are known or expected to have a rath-
er uniform distribution worldwide such as DDT,
DDE, HCB and PCBs, even when at relatively low
concentration levels. With regard to substances with
an uneven distribution in the environment and long
metabolic hls, the swift should be used with caution
as a “chemical memory” related to previous expo-
sures that may interfere with the actual measure-
ments. As exposure patterns and hls are key factors
in determining tissue levels and distribution, they
are also determinative for us to establish whether the
common swift may be suitable for use as a “sentinel
species” for air quality assessment.
It may be pointed out that the measured DDE and
PCB concentrations in the swifts studied appear, re-
Table 3 | PCDD and PCDF concentration levelsa (pg/g, fresh tissue base) in ve selected tissue types of the common swift
(sampling from April 29 to July 14, 1996)
Analyte Analyte levels per tissue type
Brain Breast muscle Heart Liver Lungs
2,3,7,8-T4CDD (D1) 0.249 0.707 0.590 1.03 0.496
1,2,3,7,8-P5CDD (D2) ≤1.05b4.15 3.54 13.8 1.24
1,2,3,4,7,8-H6CDD (D3) 0.408 5.67 7.28 39.3 5.21
1,2,3,6,7,8-H6CDD (D4) 0.840 13.7 15.3 52.0 7.31
1,2,3,7,8,9-H6CDD (D5) 0.262 4.23 4.69 23.1 3.61
1,2,3,4,6,7,8-H7CDD (D6) 0.892 9.46 13.1 150 15.2
O8CDD (D7) 5.74 20.5 64.2 641 104
2,3,7,8-T4CDF (F1) <0.40b<0.075 <0.33 0.112 0.178
1,2,3,7,8-P5CDF (F2) 0.250 0.320 <0.29 0.865 0.296
2,3,4,7,8-P5CDF (F3) 0.681 4.39 2.26 5.69 1.81
1,2,3,4,7,8-H6CDF (F4) 0.174 2.56 1.35 5.33 0.853
1,2,3,6,7,8-H6CDF (F5) 0.0929 1.39 0.860 3.24 0.638
1,2,3,7,8,9-H6CDF (F6) 0.267 1.95 1.34 2.10 1.03
2,3,4,6,7,8-H6CDF (F7) 0.211 1.95 1.16 2.95 0.877
1,2,3,4,6,7,8-H7CDF (F8) <1.0 0.493 0.535 5.44 0.523
1,2,3,4,7,8,9-H7CDF (F9) <0.17 <0.16 <0.36 0.364 <0.37
O8CDF (F10) <0.38 <0.056 <0.33 <0.19 <0.47
Totals (pg/g) 11.6 71.6 117 947 144
(a)Values corrected for analytical recovery and rounded off to three gures.
(b)Figures preceded by ≤ or < (signs respectively indicating an upper limit estimate or below limit of quantication, S/N ≈ 3; N ≈ 4 σn ) were entered as
half their nominal value to calculate cumulative data.

193
BIrds As BIoIndIcAtors of persIstent orgAnIc pollutAnts
spectively, two and one order of magnitude lower
than those reported as yielding acute effects in other
birds [32]. Therefore, focusing future research on the
role of the observed tissue concentrations in terms
of sublethal effects at population level would appear
to be appropriate, in particular to characterize the
potential for endocrine disruption. The chemicals
dealt with in this work are known to have what es-
pecially on those behavioral domains known to be
modulated by steroid hormones, such as mating,
aggression, parental care and homeostatic activities
(e.g., drinking, eating) [9].
The outcomes of our study prove that the common
swift, may be utilized as a bioindicator of environ-
mental contamination due to the chemical levels built-
up in their tissues. As mentioned in the Introduction,
behavioral studies hints also for different types of
observations [33]. In line with that, investigations car-
ried out on a swift colony in Rome revealed unexpect-
ed sexually dimorphic patterns of parental care [17,
34]. Although the authors have offered a functional
Table 4 | Selected examples of PCDD and PCDF inter-tissue variability descriptors as derived from the analytical outcomes
reported in Table 3a
Analyte Based on data sets from five tissues Based on data sets from four tissuesb
Meanc(cv%) N [Max]/[Min] Meanc(cv%) N [Max]/[Min]
D1 0.615 46.8 5 4.14 0.706 33.1 4 2.08
D2 5.68 97.7 4 11.1 5.68 97.7 4 11.1
D3 11.6 136 5 96.4 14.4 116 4 7.55
D4 17.8 112 5 61.9 22.1 91.7 4 7.12
D5 7.19 126 5 88.3 8.92 106 4 6.41
D6 37.8 167 5 168 47.0 146 4 15.9
D7 167 160 5 112 207 140 4 31.3
F1 - - 2 - - - 2 -
F2 0.433 66.9 4 3.46 0.494 65.1 3 2.92
F3 2.96 68.5 5 8.35 3.54 51.6 4 3.15
F4 2.05 98.7 5 30.6 2.52 79.3 4 6.24
F5 1.24 97.0 5 34.8 1.53 77.0 4 5.07
F6 1.34 55.4 5 7.87 1.61 31.5 4 2.03
F7 1.43 73.8 5 14.0 1.74 53.6 4 3.37
F8 1.75 141 4 11.0 1.75 141 4 11.0
F9 - - 1 - - - 1 -
F10 - - 0 - - - 0 -
Σ[PCDD+PCDF] 258 150 5 81.8 320 131 4 13.2
(a)Values appearing in the table are generally rounded off to three gures; constituents associated with gures preceded by ≤ or < (cfr. Table 3, Note b)
were not taken into account. Cfr. Table 3 for labels D1–D7 and F1–F10.
(b)Brain gures omitted.
(c)Values in pg/g, fresh tissue base.
explanation for that, we have now provided evidence
that the observed (outlying) behavioral pattern could
be a consequence of altered endocrine regulation in-
duced by the chemicals of interest. In the light of the
above, based on variation of sex typical features and
if its causal association with chemical body burden
(exposure) will be proven, this species might also nd
a future use for monitoring the exposure to – and,
therefore, the environmental occurrence of – endo-
crine disrupting agents.
Acknowledgments
We are grateful to Fabiola Ferri and Anna Maria Ingelido for her
editorial assistance, and to Giacomo Dell’Omo and Luigi Turrio
Baldassarri for their technical help in the early stages of the work.
The collection of grounded birds was allowed by the cooperation of
the Italian association for bird protection (LIPU, Section of Rome).
Claudio Carere is supported by the EU Project STARFLAG n.
12682.
Received on 2 July 2007.
Accepted on 11 March 2008.
References
1. Giesy JP, Ludwig JP, Tillitt DE. Dioxins, Dibenzofurans,
PCBs, and wildlife. In: Schecter, A (Ed.). Dioxins and health.
New York: Plenum Press; 1994. Chapter 9, p. 249-307.
2. US Environmental Protection Agency. Special report on en-
vironmental endocrine disruption: an effects assessment and
analysis. (EPA/630/R-96/012). Washington: Risk Assessment
Forum, USEPA; 1997.
3. World Health Organization. DDT and its derivatives. Geneva:
WHO; 1979. (Environmental Health Criteria, 9).
4. World Health Organization. DDT and its derivatives environ-

194 Roberto Miniero, Claudio Carere, Elena De Felip, et al.
mental aspects. International Programme on Chemical Safety.
Geneva: WHO; 1989a. (Environmental Health Criteria).
5. World Health Organization. Polychlorinated Dibenzo-para-di-
oxins and Dibenzofurans. International Programme on Chemical
Safety. Geneva: WHO; 1989b. (Environmental Health Criteria,
88).
6. World Health Organization. Polychlorinated Biphenyls and
Terphenyls (2. ed). International programme on chemical safety.
Geneva: WHO; 1993. (Environmental Health Criteria, 140).
7. International Agency for Research on Cancer. Polychlorinated
Dibenzo-para-dioxins and Polychlorinated Dibenzofurans.
Lyon: IARC; 1997. (IARC Monographs on the Evaluation
of Carcinogenic Risks to Humans, 69).
8. World Health Organization. Hexachlorobenzene. International
Programme on Chemical Safety. Geneva: WHO; 1997.
(Environmental Health Criteria, 195).
9. Dell’Omo G (Ed.). Behavioural ecotoxicology. Chichester: Wiley;
2002.
10. Dauwe T, Chu SG, Covaci A, Schepens P, Eens M. Great
tit (Parus major) nestlings as biomonitors of organochlorine
pollution. Arch Environ Contam Toxicol 2003;44:89-96.
11. Schier R, Coeurdassier M, Morilhat C, Bernard N, Faivre
B, Flicoteaux, Girardoux P, Noel M, Piotte P, Rieffel D, de
Vauleury A, Badot PM. Lead concentrations in feathers
and blood of common blackbirds (Turdus merula) and in
earthworms inhabiting unpolluted and moderately polluted
urban areas. Sci Total Environ 2006;371:197-205.
12. Angold PG, Sadler JP, Hill MO, Pullin A, Rushton S, Austin
K, Small E, Wood B, Wadsworth R, Sanderson R, Thompson
K. Biodiversity in urban habitat patches. Sci Total Environ
2006;360:196-204.
13. Sorace A. Value to wildlife of urban-agricultural parks: a case
study from Rome urban area. Environ Manag 2001;28:547-60.
14. Lack D, Owen DF. The food of the swift. J Anim Ecol 1955;
24:120-36.
15. Cramp S (Ed.). Handbook of the birds of Europe, the middle
east, and north Africa. The birds of the western palearctic. Vol.
IV. Terns to Woodpeckers. Oxford: Oxford University Press;
1985. p. 657-69.
16. Rodriguez F, Carere C, Dell’Omo G, Iacovella N, Turrio
Baldassarri L, Volpi F, di Domenico A. The common swift:
A synanthropic bird species for monitoring airborne micro-
contaminants? Organohalogen Comp 1996;28:308-13.
17. Carere C, Alleva E. Sex differences in parental care in the
common swift (Apus apus). Effect of brood size and nestling
age. Can J Zool 1998;76:1382-7.
18. Lintelmann J, Katayama A, Kurihara N, Shore L, Wenzel A.
Endocrine disruptors in the environment (IUPAC Technical
Report). Pure Appl Chem 2003;75(5)631-81.
19. Clotfelder ED, Bell AM, Leverina KR. The role of animal
behaviour in the study of endocrine-disrupting chemicals.
Animal Behav 2004;68:665-76.
20. Zala SM, Penn DJ. Abnormal behaviour induced by chemi-
cal pollution: a review of the evidence and new challenges
Animal Behav 2004;68:649-64.
21. US Environmental Protection Agency. Tetra- through octachlo-
rinated dioxins and furans by isotope dilution HRGC-HRMS.
Method 1613B. Washington: Engineering and Analysis
Division (4303), Ofce of Water, USEPA; 1994.
22. US Environmental Protection Agency. Interim procedures
for estimating risks associated with exposures to mixtures of
polychlorinated dibenzo-p-dioxins and dibenzofurans (CDDs
and CDFs) and 1989 update. Washington: Risk Assessment
Forum, USEPA; 1989. (EPA/625/3-89/016).
23. Van Leeuwen FXR, Younes MM. Consultation on assess-
ment of the health risk of dioxins: Re-evaluation of the tol-
erable daily intake (TDI). Executive summary. Food Addit
Contam 2000;17:223-40.
24. Henny CJ, Kaiser JL, Grove RA, Raymond Bentley V,
Elliot JE. Biomagnication factors (sh to osprey eggs from
Willamette River, USA) for PCDDs, PCDFs, PCBs, and OC
Pesticides. Environ Monitor Assess 2003;84:275-315.
25. Bishop CA, Koster MD, Chek AA, Hussell DJT, Jock K.
Chlorinated hydrocarbons and mercury in sediments, red-
winged blackbirds (Agelaius phoeniceus), and tree swallows
(Tachycineta bicolor) from wetlands in the Great Lakes - St.
Lawrence River basin. Environ Toxicol Chem 1995;14:491-501.
26. Wiktelius S, Edwards, CA. Organochlorine insecticide resi-
dues in African fauna: 1971-1995. Rev Environ Contam Toxicol
1997;151:1-37.
27. Mora MA. Transboundary pollution. Persistent organochlo-
rine pesticides in migrant birds of the southwestern United
States and Mexico. Environ Toxicol Chem 1997;16:3-11.
28. Turrio Baldassarri L, Carere A, di Domenico A, Fuselli S,
Iacovella N, Rodriguez F. PCDD, PCDF, and PCB contami-
nation of air and inhalable particulate in Rome. Fresenius J
Anal Chem 1994;348:144-7.
29. Berlincioni M, Croce G, Ferri F, Iacovella N, La Rocca C,
Lolini M, Megli A, Pupp M, Rizzi L, Turrio Baldassarri L,
di Domenico A. Priority organic microcontaminants in se-
lected environmental and food matrices. Fresenius Env Bull
1995;4:169-74.
30. Lack D, Lack E. The breeding behaviour of the swift. Br
Birds 1952;45:186-215.
31. Norton S. Toxic responses of the central nervous system.
In: Doull J, Klaassen CD, Amdur MO (Ed.). Casarett and
Doull’s Toxicology. The Basic Science of Poisons. New York:
Macmillan Publishing Co Inc.; 1980.
32. Noble DG, Elliott JE. Levels of contaminants in canadian
raptors, 1966 to 1988. Effects and temporal trends. Can F
Nat 1990;104:222-43.
33. Cuomo V, De Salvia MA, Petruzzi S, Alleva E. Appropriate end-
points for the characterization of behavioural changes in develop-
mental toxicology. Environ Health Perspect 1996;104:307-15.
34. Dell’Omo G, Alleva E, Carere C. Parental recycling of nes-
tling faeces in the common swift (Apus apus). Anim Behav
1998;56:631-37.