Ultrasound Obstet Gynecol 2000; 16: 918.
Fetal biometry at 4300 m compared to sea level in Peru
E. KRAMPL*, C. LEES*, J. M. BLAND, J. ESPINOZA DORADO*, G. MOSCOSO and
S. CAMPBELL
*Harris Birthright Research Center for Fetal Medicine, King's College Hospital, Denmark Hill, London SE5 9RS, UK, Department of Obstetrics
and Gynaecology and Department of Public Health Sciences, St. George's Hospital Cranmer Terrace, London SW17 0RE, UK and Instituto
Materno Perinatal, Jiron Antonio Miroquesada 149, Lima, Peru
KEY WOR DS : Fetal biometry, High altitude, Hypoxia, Ultrasound
ABSTRACT
Objectives The aim of this study was to compare
ultrasound fetal size at high altitude and sea level.
Methods Three hundred and thirty-four women in Cerro
de Pasco at 4300 m (14 100 ft) altitude and 278 women in
Lima (sea level) were recruited to the study. Ultrasound
fetal biometry was carried out between 14 and 42 weeks of
gestation. Biparietal diameter, occipitofrontal diameter,
abdominal circumference and femur length were measured
and head circumference and estimated fetal weight were
derived from these data. Two hundred and seventy-seven
women (82.9%) in Cerro de Pasco and 216 (77.7%) in
Lima had normal singleton pregnancies and certain
menstrual dates. These women were selected for statistical
analysis. Fractional polynomial regression analysis on
gestational age was performed, controlling for maternal
height and parity.
Results Fetal biometry measurements were significantly
smaller in Cerro de Pasco compared with Lima. When
gestation bands were compared this effect was present
from 25 to 29 weeks onwards, and was greater in the
abdominal circumference than in the head circumference
and femur length (ratios Cerro de Pasco : Lima, 0.96, 0.97
and 0.98, respectively). Estimated fetal weight was also
significantly lower in Cerro de Pasco (ratio 0.88), as were
birthweights (ratio 0.88). If the centiles derived from the
Lima population were applied for Cerro de Pasco, 11.2%
of all estimated fetal weights would be below the fifth
centile, and 1.08% above the ninety-fifth.
Conclusions These data suggest that at high altitude, all
fetal biometry measurements follow a lower trajectory
than at sea level. Specific biometry charts should therefore
be used for obstetric ultrasound at high altitude.
Correspondence: Elisabeth Krampl, King's College Hospital, Harris Birthright Research Center for Fetal Medicine, 9th Floor Ruskin Wing,
Denmark Hill, London SE5 9RS, UK
Received 24-11-99, Revised 23-03-00, Accepted 15-6-00
ORIGINAL ARTICLE
INTRODUCTION
Babies born at high altitude are known to be smaller than
those born at sea level. This was initially thought to be due
to a higher incidence of preterm delivery1, but the length of
gestation has subsequently been reported to be similar at
both high and low altitude26. Ascribing the difference in
birth weight simply to socio-economic differences7 has
been disproved by a large epidemiological study from
Colorado; altitude acts as an independent factor in
determining birth weight with a reduction in birth weight
of 100 g per 1000 m elevation gain6.
The smallness of babies born at high altitude is
generally attributed to lack of oxygen. It is known that
maternal diseases that reduce oxygen availability are also
associated with fetal growth restriction. For example,
mothers with cyanotic heart disease produce small
children8 and respiratory disorders adversely affect fetal
growth if maternal arterial hypoxemia is present9,10.
Animal studies have demonstrated that maternal anemia
decreases fetal growth independently from nutritional
factors11. It has been postulated that the lower birth
weight in mothers who smoke is due to the lower oxygen
availability caused by increased levels of maternal carbon
monoxide12.
At sea level, the mean barometric pressure (PB) is
760 mmHg (101.3 kPa). Oxygen comprises 21% of the air
and has a partial pressure (pO2) at sea level of 160 mmHg
(31.33 kPa); the partial oxygen pressure in alveoli (pO2alv)
is roughly 100 mmHg (13.33 kPa), arterial pO2 95 mmHg
(12.64 kPa) and hemoglobin oxygen saturation ranges
between 95 and 100%. At increasing altitude, PB falls,
pO2alv falls and the oxygen pressure difference between
alveoli and lung capillaries, which constitutes the driving
force for diffusion of gas across the alveolar membrane,
decreases. As a consequence, arterial pO2 decreases from
9
Fetal biometry at high altitude
Figure 1 Patient selection numbers in Cerro de Pasco and Lima.
95 mmHg (12.64 kPa) at sea level to about 50 mmHg
(6.67 kPa) at 4000 m and hemoglobin oxygen saturation
from 95 to 100% to about 80%.
Although an estimated 140 million people world-wide
are permanent residents at altitudes greater than
2500m13, there are no data available for ultrasound
fetal biometry at altitude. Ultrasound charts for fetal
biometry exist for Caucasians living at sea level1421, and
other ethnic groups22,23 and can be customized to take
factors such as fetal sex, maternal weight and height,
ethnic group and parity into account24. Whether low
ambient oxygen has an effect on human fetal size and the
possible magnitude of this effect has not been investigated
to date.
The aim of this study was to compare ultrasound fetal
size at high altitude and sea level in the second and third
trimesters of pregnancy in a single ethnic group. Cross-
sectional fetal biometry scans were undertaken in two
populations: in Cerro de Pasco, a mining town in the
Peruvian Andes at 4300 m (14 100 ft), and in Lima (sea
level).
SUBJECTS AND METHODS
The study was carried out as a collaboration between
academic obstetric departments in London, UK and Lima,
Peru. The ethics committee of the Peruvian Ministry of
Health, and the IPSS (Instituto Peruano de Seguro Social)
hospital manager of the Pasco district gave ethical
approval. All subjects gave informed written consent for
ultrasound examination having read an information sheet
in Spanish. In addition, a local research nurse verbally
explained the purpose of the study.
Subjects were recruited both in Lima and in Cerro de
10
Krampl et al.
Pasco from hospital antenatal and health clinics, when they
attended for routine visits. Local physicians, midwives and
nursing students carried out the recruitment. They had
been given detailed information about the study by the
examiners and had attended the examinations. All women
were ethnic Mestizo, a mixture between native Quechua
and Spanish. Only women who were resident at the
altitude at which they were born were included in the
study.
One operator (E.K.) performed ultrasound fetal biome-
try using Aloka SSD-1700 (Aloka Co., Ltd, Tokyo, Japan)
and Toshiba Eccocee (Toshiba, Japan) Ultrasound Appa-
ratus using 3.5 and 5 MHz probes. Biometry measure-
ments were stored as thermal printed images with
individual patient records. Measurements of biparietal
diameter (BPD), occipitofrontal diameter (OFD) and
abdominal circumference (AC) were all made in the planes
described previously25, where BPD and OFD were measured
from outer to outer border of the skull in a plane
demonstrating the anterior and posterior horns of the lateral
ventricles and AC was measured using an ellipse. For the
femur length (FL) a straight measurement was taken from
the greater trochanter to the lateral condyle in a plane at
approximately 458 to the ultrasound beam21. Head circum-
ference (HC) was calculated from 3.14  (BPD 1 OFD)/2.
Fetal weight was estimated (EFW) according to Hadlock
(HC 2 BPD 2 AC 2 FL)26. The following ratios were
calculated: HC/AC, HC/FL and AC/FL.
Healthy women referred with normal singleton preg-
nancies with a gestational age of 14 weeks or more were
examined once. Patients referred with a known clinical
problem which was likely to influence fetal growth, such as
hypertension, diabetes or clinical suspicion of intra-uterine
growth restriction were not included in the study. A total
number of 612 scans was performed, 334 in Cerro de
Pasco and in 278 in Lima (Figure 1). Women who were not
sure of their last menstrual period (LMP), did not
remember it, who had irregular periods or stopped taking
the contraceptive pill less than three cycles before
becoming pregnant, or where there were multiple preg-
nancies and fetal malformations were excluded. When the
biometry measurements (282 from Cerro de Pasco and 218
from Lima) were plotted against the gestational age it was
evident that five cases (1.5%) from Cerro de Pasco and two
cases (0.7%) from Lima were extreme outliers. Their
biometry measurements (BPD, OFD, AC, FL) prior to
25 weeks differed by more than 4 weeks from the mean of
gestation. It was assumed that this was due to error in
maternal recall of the menstrual history and as the
intention was to establish normal ranges for fetal biometry
at a given gestation these measurements were excluded
from the analysis (Figure 2). Statistical analysis was
performed on the measurements of the remaining 493
women (277 from Cerro de Pasco and 216 from Lima)
(Figure 1).
In the first 2 weeks of the study, pO2 and pCO2 were
measured in all patients with a portable blood gas analyzer
(OPTI 1, AVL, Graz, Austria) using solid-state single use
optical fluorescence cassettes. Full 1point gas calibration
Ultrasound in Obstetrics and Gynecology
Fetal biometry at high altitude
Figure 2 Head circumference measurements in Cerro de Pasco and
Lima. A total of 502 measurements, 218 from Lima and 284 from
Cerro de Pasco and from women with normal singleton pregnancies
with reliable menstrual history are plotted by gestational age derived
from the date of last menstrual period (LMP). Triangles denote five
patients from Cerro de Pasco and two patients from Lima who were
excluded from analysis because of maternal error in recall of the
LMP.
for pO2 and pCO2 was carried out automatically after
insertion of each cassette. Stable standard reference
cassettes were used for verification of low, medium and
high levels. Arterialized capillary samples were taken from
the earlobe. Samples obtained by this method have been
shown to have blood gas contents similar to arterial
samples27. Maternal hematocrit was determined by the
microcapillary method28.
Statistical analysis
Statistical analysis was performed on the measurements of
the remaining 493 women (277 from Cerro de Pasco and
216 from Lima). The planned sample size was 200 in each
town. The standard error of the sample size can be
established using the method of Altman and Chitty29. For a
sample size of 200 the width of the 95% confidence
interval for the tenth and ninetieth centiles would
be ^ 0.19 standard deviations. As we exceeded 200 in
Figure 3 Normal distribution of the abdominal circumference (AC) measurements corrected for altitude: (a), frequency distribution of the resi-
duals; (b), residuals plotted against gestational age (weeks).
Ultrasound in Obstetrics and Gynecology
Krampl et al.
each town our estimates are rather more precise than this.
Gestational age was adjusted by using the fractional
polynomial method of Royston and Altman30. This allows
a complex curve to be fitted using a small number of
parameters, enabling the effect of gestational age to be
removed more efficiently than could be done using a simple
polynomial. The variability of each biometric measure-
ment clearly increased as the gestation advanced. The
standard deviation appeared to increase with increasing
mean measurement. To make the variability independent
of the magnitude and hence suitable for regression
analysis, the data were log-transformed. Biometric vari-
ables were regressed on gestational age, altitude (Lima 0,
Cerro de Pasco 1), parity and maternal height. These
variables were specified in advance, no stepwise procedure
was used. The coefficient of altitude estimated the
difference between Cerro de Pasco and Lima. Anti-log
transformation of this coefficient yielded the Cerro de
Pasco/Lima ratio. This adjusted ratio is the estimate of the
size of a fetus in Cerro de Pasco divided by the size of the
fetus in Lima as if they were of the same gestational age
and maternal height. The regression method provided a
confidence interval for this ratio and a test of the null
hypothesis that the population ratio equaled 1.
The residuals (the difference between the observed
biometric measurement and that predicted by the regres-
sion equation) were calculated. The mean of the absolute
value of the residual, irrespective of sign, can be used to
estimate the standard deviation provided the underlying
distribution is normal31. The absolute residuals were then
regressed on gestational age, again using the fractional
polynomial method. The mean of the half-normal dis-
p
tribution is 2 over p. Multiplying the predicted absolute
p
residual by p over 2 thus gives the standard deviation at
this gestation. The mean and standard deviation were used
to calculate centiles using the normal distribution.
Interaction between the population and gestational age
was calculated by adding a gestational age  group term
to the regression model. Interaction would occur if the
relative difference between the Cerro de Pasco and Lima
populations were not the same at each gestational age. An
altitude effect threshold model was also developed. A series
11
Fetal biometry at high altitude
Figure 4 Fiftieth, fifth and ninety-fifth centile for biparietal diameter
(BPD), derived from 493 measured fetuses, 216 from Lima and 277
from Cerro de Pasco.
of variables was defined, each of which was always zero
for fetuses from Cerro de Pasco. For fetuses from Lima, the
variable was zero unless the gestational age was within a
certain range, usually of 5 weeks (1519, 2024, 2529.
3034, 3540). Regression analysis with these variables as
predictors was then performed. This allowed different
slopes to be estimated for each interval of gestational age.
If there were a threshold effect, these variables would only
affect biometry at later gestational ages.
Figure 3 shows the distribution of the residual abdom-
inal circumference after log transformation and fitting
gestational age and town effects. The variability appears
uniform throughout the range of gestational age and the
shape of the distribution is symmetrical, fitting the normal
quite closely. Similar results were found for all the
biometry measurements. Normal regression methods can
therefore be used on the log-transformed data. The
abdominal circumference was chosen for graphical illus-
tration because pathological growth restriction affects the
abdominal circumference more than other biometry and
would therefore skew the distribution the most.
The reliability of measurements was assessed on 62 sets of
Figure 6 Fiftieth, fifth and ninety-fifth centile for abdominal circum-
ference (AC), derived from 493 measured fetuses, 216 from Lima
and 277 from Cerro de Pasco.
12
Krampl et al.
Figure 5 Fiftieth, fifth and ninety-fifth centile for head circumfer-
ence (HC), derived from 493 measured fetuses, 216 from Lima and
277 from Cerro de Pasco.
repeated measurements taken in Cerro de Pasco using the
intra-class correlation coefficient32. This is the correlation
between repeated measurements ignoring the ordering of the
pair. All intra-class correlations were between 0.982 and
0.997.
The t-test was used to compare maternal characteristics
between Cerro de Pasco and Lima, except for parity (Mann
Whitney U-test) and the level of education (x2 test).
Analyses were performed using Stata 5.0 (Stata Cor-
poration, TX, USA) and SPSS 8.0 (SPSS Science, Chicago,
IL, USA).
R ES U LT S
Table 1 shows that maternal age and pre pregnancy weight
were similar in Cerro de Pasco and Lima. Height was
however, significantly less in Cerro de Pasco compared
with Lima (149.9 ^ 5.3 cm versus 152.8 ^ 6.0 cm,
P , 0.001) and parity significantly higher (1.21 ^ 1.6
versus 0.64 ^ 1.1). Maternal levels of pO2 and pCO2 were
significantly lower in Cerro de Pasco than in Lima
(52.8 ^ 3.5 mmHg versus 101.6 ^ 14 mmHg, P , 0.001
Figure 7 Fiftieth, fifth and ninety-fifth centile for femur length (FL),
derived from 493 measured fetuses, 216 from Lima and 277 from
Cerro de Pasco.
Ultrasound in Obstetrics and Gynecology
Fetal biometry at high altitude
Figure 8 Fiftieth, fifth and ninety-fifth centile for estimated fetal
weight (EFW), derived from 493 measured fetuses, 216 from Lima
and 277 from Cerro de Pasco.
and 25.5 ^ 2.4 mmHg versus 38.0 ^ 12.3 mmHg,
P , 0.001, respectively). Maternal hematocrit was sig-
nificantly higher (43.9 ^ 8.1% versus 36.5 ^ 3.5%
P , 0.001) (Table 1). The socio-economic status as
assessed by the level of school education was similar in
both populations. In Cerro de Pasco 211 women (76.2%)
had completed secondary school, in Lima the number was
170 (78.7%).
Fetal biometric measurements (BPD, OFD, HC, AC and
FL) from Cerro de Pasco were compared with Lima by
regression analysis controlling for gestational age, mater-
nal height and parity (Tables 29), (Figures 47). There
was a statistically significant difference between Cerro de
Pasco and Lima. The ratios Cerro de Pasco/Lima for the
five variables were 0.98, 0.97, 0.97, 0.96, 0.98, respec-
tively (Table 10). The EFW was also significantly lower in
Cerro de Pasco [ratio 0.88, 95% confidence interval (CI)
0.840.91)] (Figure 8). Birth weights were available from
360 live births at term, 213 (76.9%) from Cerro de Pasco
and 147 (68.1%) from Lima. The mean uncorrected
birthweights were 2944 g and 3351 g, respectively, the
ratio Cerro de Pasco/Lima being 0.88 (95% CI 0.850.90).
The ratio HC/AC was significantly larger in Cerro de
Pasco than in Lima (ratio 1.01, P 0.003) (Figure 9), the
ratio AC/FL was significantly smaller (ratio 0.98,
P 0.01), but the ratio HC/FL was similar. These effects
did not change with gestation (Table 10).
There was significant interaction between group and
gestational age with respect to the BPD, OFD, HC, AC and
FL measurements (P-values 0.002, 0.006, 0.02, , 0.001
and 0.005, respectively). The threshold model showed that
the difference becomes apparent from the gestational age
interval 2529 weeks onwards (Table 11).
If the centiles derived from Lima are applied to the Cerro
de Pasco population, 11.2% of EFWs over all gestation
ranges are below the fifth centile, and 1.08% above the
ninety-fifth. The effect is more pronounced from 35 weeks
of gestation onwards with 21% of the EFWs at Cerro de
Pasco being below the Lima fifth centile and none above
the ninety-fifth (Table 12).
The regression equations are summarized in Table 13.
Ultrasound in Obstetrics and Gynecology
Krampl et al.
Figure 9 Fiftieth, fifth and ninety-fifth centile for ratio of head cir-
cumference to abdominal circumference (HC/AC), derived from 493
measured fetuses, 216 from Lima and 277 from Cerro de Pasco.
DISCUSSION
We present the first comparative ultrasound biometry data
for high altitude and sea level. All fetal measurements
proved to be smaller in Cerro de Pasco than in Lima. When
gestation bands were compared this effect was present
from 25 to 29 weeks onwards. These charts are likely to
reflect fetal growth in utero accurately as the ratio Cerro de
Pasco/Lima between the EFW and the actual birth weights
was very similar. Furthermore, all ultrasound measure-
ments showed a very high degree of reliability as measured
by intraclass correlation. The reference ranges are reliable
because the sample size was adequate, the measurements
were performed in ethnically similar groups with similar
socio-economic status and appropriate adjustments were
made for important predictors of birth weight such as
maternal height and parity. We therefore interpret the
difference in fetal size between Cerro de Pasco and Lima as
being due to altitude.
A major difference between the studied population and
Europeans33 was that it is highly unusual for pregnant
women in Peru to smoke; there were no smokers among
the women examined. Another interesting difference was
the high recall of LMP in this population both in Lima and
in Cerro de Pasco. Around 80% were sure of their LMP,
compared with 55% of the population scanned routinely at
a London teaching hospital34. This may partly reflect the
lower incidence of oral contraceptive use, but may also be
related to the higher motivation of women volunteering to
take part in a research study. The level of education was
high in both populations, even higher in Cerro de Pasco;
the charts shown may arguably be biased against those at
the lowest end of the socio-economic spectrum.
It has also become evident that the AC is disproportio-
nately smaller than HC and FL. This phenomenon has
been shown previously in animal studies: pregnant rabbits
placed in a hypobaric chamber to simulate an altitude of
12 000 ft (3650 m) have smaller offspring and the size of
their organs is consistent with asymmetrical growth
restriction; the brain weight was similar, whereas the
weight of the liver was significantly smaller in those
13
Fetal biometry at high altitude Krampl et al.

Table 1 Maternal clinical data; *n 67 in Cerro de Pasco, n 40 Table 3 Reference table for biparietal diameter in Cerro de Pasco
in Lima
GA (weeks) 5th centile Mean 95th centile
Cerro de Pasco Lima
Mean (SD) Mean (SD) P-value 14 22.9 26.4 29.8
15 25.9 29.9 33.9
n (analyzed) 277 216 16 29.0 33.5 38.0
Age 27.4 (5.9) 26.3 (6.9) NS 17 32.2 37.0 41.9
Parity 1.2 (1.5) 0.6 (1.1) , 0.001 18 35.4 40.6 45.7
Weight (kg) 54.2 (6.9) 55.8 (8.7) NS 19 38.6 44.0 49.5
Height (cm) 149.9 (5.3) 152.8 (6.0) , 0.001 20 41.8 47.5 53.1
pO2 (mmHg)* 52.8 (3.5) 101.6 (14.0) , 0.001 21 45.0 50.8 56.7
pCO2 (mmHg)* 25.5 (2.4) 38.0 (12.3) , 0.001 22 48.1 54.1 60.1
Hematocrit * 43.9 (8.1) 36.5 (3.5) , 0.001 23 51.1 57.3 63.5
24 54.1 60.4 66.7
25 56.9 63.4 69.8
26 59.6 66.2 72.8
rabbit offspring that developed under hypobaric condi- 27 62.3 69.0 75.6
tions35. 28 64.8 71.6 78.3
29 67.1 74.0 80.9
We assessed maternal blood gas levels in a subgroup of
30 69.3 76.3 83.3
women participating in this study. Maternal pO2 in Cerro 31 71.4 78.4 85.5
de Pasco was only 51% of that in Lima. The levels of 32 73.2 80.4 87.5
maternal pCO2 in Cerro de Pasco are 66% of those in 33 74.9 82.2 89.4
Lima suggesting that this relative hypocapnia can be 34 76.5 83.8 91.0
35 77.8 85.2 92.5
explained by a compensatory increase in maternal
36 78.9 86.4 93.8
ventilation. This is known to occur in non-pregnant 37 79.8 87.4 94.9
humans36. Furthermore, hypoxia is known to induce the 38 80.6 88.1 95.7
erythropoietin gene37 and increase hematocrit38; our data 39 81.0 88.7 96.4
for maternal hematocrit confirm this effect. These 40 81.3 89.0 96.8
compensatory mechanisms in the mother do not seem to GA, gestational age.
be sufficient to maintain oxygen availability for the feto-
placental unit. At altitude the placenta is larger and there

Table 2 Reference table for biparietal diameter in Lima Table 4 Reference table for head circumference in Lima

GA (weeks) 5th centile Mean 95th centile GA (weeks) 5th centile Mean 95th centile

14 22.4 24.9 27.3 14 75.0 87.6 100.2

15 25.3 28.7 32.0 15 87.5 101.7 115.8
16 28.4 32.5 36.6 16 100.2 115.7 131.2
17 31.6 36.3 41.1 17 112.9 129.7 146.5
18 34.8 40.1 45.4 18 125.6 143.5 161.4
19 38.1 43.8 49.6 19 138.3 157.2 176.1
20 41.4 47.5 53.6 20 150.9 170.6 190.4
21 44.7 51.1 57.5 21 163.3 183.8 204.3
22 48.0 54.6 61.3 22 175.5 196.6 217.7
23 51.2 58.1 64.9 23 187.4 209.0 230.7
24 54.4 61.4 68.4 24 199.0 221.1 243.2
25 57.5 64.6 71.6 25 210.2 232.6 255.1
26 60.5 67.6 74.7 26 221.0 243.7 266.4
27 63.4 70.5 77.7 27 231.4 254.2 277.1
28 66.2 73.3 80.4 28 241.3 264.2 287.1
29 68.9 75.9 82.9 29 250.7 273.6 296.5
30 71.5 78.4 85.3 30 259.5 282.3 305.1
31 73.9 80.6 87.4 31 267.8 290.4 313.0
32 76.1 82.7 89.3 32 275.4 297.8 320.1
33 78.2 84.6 91.0 33 282.3 304.4 326.4
34 80.1 86.3 92.4 34 288.6 310.3 331.9
35 81.8 87.7 93.7 35 294.2 315.4 336.6
36 83.4 89.0 94.7 36 299.0 319.6 340.3
37 84.7 90.0 95.4 37 303.0 323.1 343.2
38 85.8 90.9 95.9 38 306.2 325.6 345.1
39 86.7 91.4 96.1 39 308.6 327.3 346.0
40 87.4 91.7 96.1 40 310.1 328.1 346.0

GA, gestational age. GA, gestational age.

14 Ultrasound in Obstetrics and Gynecology

Fetal biometry at high altitude Krampl et al.

Table 5 Reference table for head circumference in Cerro de Pasco Table 7 Reference table for abdominal circumference in Cerro de
Pasco
GA (weeks) 5th centile Mean 95th centile
GA (weeks) 5th centile Mean 95th centile
14 80.0 91.0 102.1
15 90.4 104.4 118.3 14 68.4 82.7 96.9
16 101.5 117.7 133.9 15 78.6 93.6 108.7
17 113.1 130.9 148.8 16 88.9 104.7 120.6
18 124.8 144.1 163.3 17 99.3 115.9 132.5
19 136.7 157.0 177.3 18 109.8 127.1 144.4
20 148.5 169.7 190.9 19 120.3 138.3 156.3
21 160.3 182.1 204.0 20 130.7 149.4 168.2
22 171.8 194.2 216.6 21 141.0 160.5 180.1
23 183.1 206.0 228.8 22 151.2 171.5 191.9
24 194.1 217.3 240.6 23 161.2 182.4 203.6
25 204.7 228.2 251.8 24 171.0 193.1 215.2
26 214.8 238.6 262.4 25 180.5 203.6 226.7
27 224.5 248.5 272.5 26 189.8 213.9 238.0
28 233.7 257.9 282.1 27 198.8 224.0 249.2
29 242.3 266.6 291.0 28 207.5 233.8 260.2
30 250.4 274.8 299.2 29 215.8 243.4 270.9
31 257.8 282.3 306.8 30 223.7 252.6 281.5
32 264.5 289.1 313.8 31 231.2 261.5 291.8
33 270.6 295.3 319.9 32 238.3 270.1 301.8
34 276.0 300.7 325.4 33 245.0 278.3 311.6
35 280.6 305.3 330.1 34 251.2 286.1 321.1
36 284.4 309.2 334.0 35 256.9 293.6 330.2
37 287.4 312.2 337.1 36 262.0 300.6 339.1
38 289.6 314.4 339.3 37 266.7 307.1 347.6
39 290.9 315.8 340.7 38 270.7 313.2 355.8
40 291.4 316.3 341.1 39 274.2 318.9 363.6
40 277.1 324.1 371.0
GA, gestational age.
GA, gestational age.

Table 6 Reference table for abdominal circumference in Lima Table 8 Reference table for femur length in Lima

GA (weeks) 5th centile Mean 95th centile GA (weeks) 5th centile Mean 95th centile

14 60.3 71.6 82.9 14 6.4 10.2 14.0

15 72.7 85.2 97.7 15 10.2 14.1 18.0
16 84.8 98.7 112.5 16 13.8 17.8 21.8
17 96.7 112.0 127.2 17 17.3 21.4 25.5
18 108.4 125.1 141.7 18 20.6 24.8 29.0
19 119.9 138.0 156.1 19 23.8 28.1 32.3
20 131.1 150.7 170.3 20 26.9 31.2 35.6
21 142.0 163.2 184.4 21 29.9 34.3 38.7
22 152.7 175.5 198.2 22 32.7 37.2 41.7
23 163.2 187.5 211.8 23 35.4 40.0 44.5
24 173.4 199.3 225.1 24 38.0 42.7 47.3
25 183.4 210.8 238.2 25 40.5 45.2 49.9
26 193.2 222.0 250.9 26 42.9 47.7 52.5
27 202.7 233.0 263.3 27 45.2 50.1 54.9
28 212.0 243.7 275.4 28 47.3 52.3 57.3
29 221.1 254.1 287.1 29 49.4 54.5 59.5
30 230.0 264.1 298.3 30 51.3 56.5 61.7
31 238.6 273.9 309.1 31 53.2 58.4 63.7
32 247.1 283.3 319.5 32 54.9 60.3 65.7
33 255.3 292.3 329.3 33 56.5 62.0 67.5
34 263.4 301.0 338.7 34 58.0 63.6 69.3
35 271.3 309.4 347.5 35 59.4 65.2 70.9
36 279.0 317.3 355.7 36 60.7 66.6 72.5
37 286.5 324.9 363.3 37 61.8 67.9 74.0
38 293.9 332.1 370.2 38 62.9 69.1 75.4
39 301.1 338.8 376.5 39 63.9 70.3 76.7
40 308.2 345.2 382.1 40 64.7 71.3 77.8

GA, gestational age. GA, gestational age.

Ultrasound in Obstetrics and Gynecology 15

Fetal biometry at high altitude Krampl et al.

Table 9 Reference table for femur length in Cerro de Pasco Table 11 Threshold model. The threshold model was applied to all
fetal biometric measurements with significant interaction. The ratio
GA (weeks) 5th centile Mean 95th centile Cerro de Pasco/Lima is consistently smaller at later gestations than at
earlier. The differences become statistically significant from 25 to
14 8.8 12.7 16.5 29 weeks onwards
15 11.7 15.8 19.9
16 14.6 18.9 23.2 Interaction Gestational age interval (weeks)
17 17.5 21.9 26.4
18 20.3 24.9 29.5 Variable P-value , 20 2024 2529 3034 $ 35
19 23.1 27.8 32.6
20 25.9 30.7 35.5 BPD 0.002 1.00 1.01 0.97 0.97 0.97
21 28.5 33.5 38.4 OFD 0.006 1.00 1.00 0.98 0.96 0.96
22 31.2 36.2 41.2 HC 0.02 0.99 1.00 0.97 0.96 0.96
23 33.7 38.8 43.9 AC , 0.001 1.02 0.99 0.96 0.94 0.94
24 36.2 41.4 46.6 FL 0.005 1.03 0.99 0.98 0.95 0.97
25 38.6 43.9 49.1
26 40.9 46.2 51.6 BPD, biparietal diameter, OFD, occipitofrontal diameter; HC, head
27 43.1 48.5 54.0 circumference; AC, abdominal circumference; FL, femur length.
28 45.2 50.7 56.3
29 47.2 52.8 58.5
30 49.1 54.8 60.6 at high altitude are normally distributed and not skewed
31 50.9 56.7 62.5 towards smaller measurements. Furthermore, the neonatal
32 52.7 58.5 64.4 and infant mortality rates in low birthweight infants in
33 54.2 60.2 66.2
34 55.7 61.8 67.9
high-altitude populations have been shown to be lower
35 57.1 63.2 69.4 compared to low altitudes in Mexico42, Peru43 and
36 58.3 64.6 70.9 Colorado44.
37 59.4 65.8 72.2 The mechanisms which constrain fetal growth to
38 60.4 66.9 73.4 enable adaptation to altitude are poorly understood.
39 61.2 67.8 74.4
40 61.9 68.6 75.4
There is no doubt that maternal hypoxia predisposes to
small fetal size. In humans at high altitude, maternal
arterial oxygenation during pregnancy45 and the mater-
nal ventilatory response to hypoxic stress is directly
is increased branching of the capillaries with a decreased correlated with birth weight46. Further work might
diffusion distance at high altitude39,40. The fetus shows address non-invasive means of assessing fetal oxygena-
increased erythropoiesis: babies born at a similar high tion such as Doppler investigation of the arterial and
altitude in Bolivia (3800 m, 12 500 ft) have been shown venous circulations, and maternal physiologic and endo-
to have a higher hematocrit, higher levels of hemoglobin crine changes that occur throughout pregnancy at high
and fetal hemoglobin in cord blood compared with sea altitude.
level41. Importantly our results suggest that before 25 weeks,
Lower average birthweight at high altitude is likely to be fetal size at high altitude is not significantly different from
due to physiological adaptation rather than a higher sea level. Therefore second trimester dating charts from sea
number of pathologically growth-restricted babies for the
following reasons. Our data show that the measurements
Table 12 The percentage of measurements from Cerro de Pasco that
are below the fifth centile and above the ninety-fifth centile derived
from the Lima population
Table 10 Regression analysis of biometric variables controlled for
maternal height and parity Gestational age interval (weeks)

Ratio 95% confidence Variable , 25 2529 3034 35 1 Total

Variable Cerro de Pasco/Lima P-value interval Interaction
BPD , 5th 6.06 9.09 15.62 23.46 14.08
BPD 0.98 0.002 0.960.99 0.002 . 95th 9.09 3.03 3.12 2.47 4.33
OFD 0.97 0.001 0.960.99 0.006 OFD , 5th 6.06 10.61 14.06 19.75 13.00
HC 0.97 , 0.001 0.960.98 0.02 . 95th 3.03 4.55 7.81 1.23 3.97
AC 0.96 , 0.001 0.940.97 , 0.001 HC , 5th 7.58 10.61 14.06 23.46 14.44
FL 0.98 0.02 0.960.99 0.005 . 95th 6.06 4.55 4.69 0.00 3.61
EFW 0.88 , 0.001 0.840.91 0.2 AC , 5th 4.55 10.61 6.25 22.22 11.55
HC/AC 1.01 0.003 1.001.02 0.2 . 95th 7.58 3.03 1.56 0.00 2.89
HC/FL 1.00 0.9 0.981.01 0.004 FL , 5th 10.61 10.61 9.38 13.58 11.19
AC/FL 0.98 0.01 0.970.99 0.01 . 95th 7.58 7.58 0.00 1.23 3.97
EFW , 5th 3.03 7.58 10.94 20.99 11.19
BPD, biparietal diameter, OFD, occipitofrontal diameter; HC, head . 95th 0.00 3.03 1.56 0.00 1.08
circumference; AC, abdominal circumference; FL, femur length; EFW,
estimated fetal weight; HC/AC, head circumference/abdominal BPD, biparietal diameter; OFD, occipitofrontal diameter; HC, head
circumference ratio; HC/FL, head circumference/femur length ratio; circumference; AC, abdominal circumference; FL, femur length; EFW,
AC/FL, abdominal circumference/femur length ratio. estimated fetal weight.

16 Ultrasound in Obstetrics and Gynecology

Fetal biometry at high altitude
Table 13 Regression equations
Cerro de Pasco
BPD
OFD
HC
AC
FL
EFW
Lima
BPD
OFD
HC
AC
FL
EFW
BPD, biparietal diameter; OFD, occipitofrontal diameter; HC, head circumference; AC,
abdominal circumference; FL, femur length; EFW, estimated fetal weight; gest, gestational age;
sqrt, square root.
level may be valid at high altitude. On the contrary, from
35 weeks onwards, 21% of the fetuses in Cerro de Pasco
have an estimated fetal weight below the fifth centile of the
Lima population, and none are above the ninety-fifth
centile. The use of inappropriate charts of fetal biometry in
high altitude settings would lead to an over-diagnosis of
small-for-gestational age babies, and the under-diagnosis of
potential macrosomic conditions. Specific biometry charts
should therefore be used for obstetric ultrasound at high
altitude.
ACKNOWLEDGEMENTS
We would like to thank Dr Pedro Mascaro Sanchez,
Director, and Dr Gloria Larrabure, Research Director,
from the Instituto Materno-Perinatal in Lima and Dr
Mario Soto, Hospital Manager and Gloria Castro, Head
of Midwifery, from Cerro de Pasco. Nilda Janampa, the
Research Nurse, was extremely committed and her work
was essential for running the clinics.
We are grateful to Toshiba UK, Aloka Peru and AVL
Austria for providing the equipment and the Peruvian
Ambassador in London, HE J. Eduardo Ponce Vivanco and
his staff for their invaluable support.
This project was supported by the Fetal Medicine
Foundation, London (Registered Charity 1037116), and
E.K. was supported by the Austrian Science Foundation,
FWF (Erwin Schrodinger Fellowship, registered project
number J01625).
REFERENCES
1 Lichty JA, Ting RY, Bruns PD, Dyar E, Incidence of prematurity
higher at high altitude. Public Health Rep 1955; 70: 230.
2 Lichty JA, Ting RY, Bruns PD. Studies of babies born at high
altitude I. Relation of altitude to birth weight. Am J Dis Child
1957; 93: 6669
3 McClung J. Effect of High Altitude on Human Birth. Harvard,
MA: Harvard University Press, 1969
4 Kruger H, Arias-Stella J. The placenta and the newborn infant at
high altitudes. Am J Obstet Gynecol 1970; 106: 58691
5 Haas JD. Prenatal and infant growth and development. In: Baker
Ultrasound in Obstetrics and Gynecology
Krampl et al.
2 10.11463 1 22.59607  gest2 2 11.8292  gest2  log(gest)
2 18.32153 1 31.38366  gest2 2 16.73594  gest2  log(gest)
2 46.63544 1 85.48547  gest2 2 45.3033  gest2  log(gest)
2 24.37415 1 65.13188  gest2 2 31.27398  gest2  log(gest)
2 34.88533 1 35.09442  gest 2 05760241  gest3
6933.429 1 13628.24  gest22 2 57328.36  gest22  log(gest)
2 14.43395 1 24.34002  gest2 2 12.77063  gest2  log(gest)
2 22.17255 1 32.92824  gest2 2 17.49558  gest2  log(gest)
2 57.20661 1 89.83847  gest2 2 47.4349  gest2  log(gest)
2 130.2058 1 147.6567  gest 2 1.800975  gest3
2 104.8855 1 98.01394  sqrt(gest) 2 0.3105365  gest3
12842.6 1 52806.21  gest22 2 51276.29  gest21
PT, Little MA, eds. Man in the Andes, 1st edn. Stroudsburg, PA:
Dowdon, Hutchinson & Ross, 1976: 16179
6 Jensen GM, Moore LG. The effect of high altitude and other risk
factors on birthweight: independent or interactive effects? Am J
Public Health 1997; 87: 10037
7 Cotton EK, Hiestand M, Philbin GE, Simmons M. Re-evaluation
of birth weights at high altitude. Study of babies born to mothers
living at an altitude of 3,100 meters. Am J Obstet Gynecol 1980;
138: 2202.
8 Cannell DE, Vernon CP. Congenital heart disease and pregnancy.
Am J Obstet Gynecol 1963; 85: 744
9 De Swiet M. Respiratory disease in pregnancy. Postgrad Med J
1979; 55: 3258
10 Schatz M, Zeiger RS, Hoffman CP. Intrauterine growth is related
to gestational pulmonary function in pregnant asthmatic women.
Kaiser-Permanente Asthma and Pregnancy Study Group. Chest
1990; 98: 38992
11 Crowe C, Dandekar P, Fox M, Dhingra K, Bennet L, Hanson MA.
The effects of anemia on heart, placenta and body weight, and
blood pressure in fetal and neonatal rats. J Physiol (Lond) 1995;
488: 5159
12 Bureau MA, Shapcott D, Berthiaume Y, Monette J, Blouin D,
Blanchard P, Begin R, Bureau MA, Shapcott D, Berthiaume Y,
Monette J, Blouin D, Blanchard P, Begin R. Maternal cigarette
smoking and fetal oxygen transport: a study of P50, 2,3-
diphosphoglycerate, total hemoglobin, hematocrit, and type F
hemoglobin in fetal blood. Pediatrics 1983; 72: 226
13 Moore LG, Niermeyer S, Zamudio S. Human adaptation to high
altitude: regional and life-cycle perspectives. Am J Phys Anthropol
Suppl 1998; 27: 2564
14 Campbell S, Newman GB. Growth of the fetal biparietal diameter
during normal pregnancy. J Obstet Gynaecol Br Commonwealth
1971; 78: 5139
15 Hadlock FP, Harrist RB, Russsell LD, Park SK. Fetal head
circumference. Am J Radiol 1982; 138: 64953
16 Hadlock FP, Harrist RB, Russsell LD, Park SK. Fetal abdominal
circumference. Am J Radiol 1982; 139: 36770
17 Hadlock FP, Harrist RB, Russsell LD, Park SK. Fetal femur length
as a predictor of menstrual age. Am J Radiol 1982; 138: 8758
18 Chitty LS, Altman DG, Henderson A, Campbell S. Charts of fetal
size: 2. Head measurements. Br J Obstet Gynaecol 1994; 101: 35
43
19 Chitty LS, Altman DG, Henderson A, Campbell S. Charts of fetal
size: 3. Abdominal measurements. Br J Obstet Gynaecol 1994;
101: 12531
20 Chitty LS, Altman DG, Henderson A, Campbell S. Charts of fetal
size: 4. Femur length. Br J Obstet Gynaecol 1994; 101: 1325
21 Snijders RJM, Nicolaides KH. Fetal biometry at 1440 weeks'
gestation. Ultrasound Obstet Gynecol 1994; 4: 3848
17
Fetal biometry at high altitude
22 Browne PC, Hamner LH, Clark WS. Sonographic fetal growth
curves from an indigent population in Atlanta, Georgia. I.
Singleton pregnancies. Am J Perinatol 1992; 9: 46776
23 Davis RO, Cutter GR, Goldenberg RL, Hoffman HJ, Cliver SP,
Brumfield CG. Fetal biparietal diameter, head circumference,
abdominal circumference and femur length. A comparison by race
and sex. J Reprod Med 1993; 38: 2016
24 Gardosi J, Chang A, Kalyan B, Sahota D, Symonds EM.
Customized antenatal growth charts. Lancet 1992; 339: 2837
25 Campbell S, Thoms A. Ultrasound measurement of the fetal head
to abdomen circumference ratio in the assessment of growth
retardation. Br J Obstet Gynaecol 1977; 84: 16574
26 Hadlock FP, Harrist RB, Sharman RS, Deter RL, Park SK.
Estimation of fetal weight with the use of head, body, and femur
measurements a prospective study. Am J Obstet Gynecol 1985;
151: 3337
27 White RD, Hoopes KH, Allen SD. Comparison of capillary ear
blood and -arterial blood to validate capillary sampling as an
accurate assay of blood gas in swine. Am J Vet Res 1979; 40:
10224
28 Platt WR, ed. Color Atlas and Textbook of Hematology
Philadelphia: JB Lippincott, 1969: 325.
29 Altman DG, Chitty LS. Charts of fetal size: methodology. Br J
Obstet Gynaecol 1994; 101: 2934
30 Royston P, Altman DG. Regression using fractional polynomials
or continuous covariates parsimonious parametric modeling.
Appl Stat J Royal Stat Soc Series C 1994; 43: 1214.
31 Royston P, Altman DG. Construction of age-related reference
centiles using absolute residuals. Stat Med 1993; 12: 91724
32 Bland JM, Altman DG. Measurement error and correlation
coefficients. BMJ 1996; 313: 412
33 Graham H. Smoking prevalence among women in the European
Community 195090 Soc Sci Med 1996; 43: 24354
34 Campbell S, Warsof SL, Little D, Cooper DJ. Routine ultrasound
screening for the prediction of gestational age. Obstet Gynecol
1985; 65: 61320
18
Krampl et al.
35 Chang JH, Rutledge JC, Stoops D, Abbe R. Hypobaric hypoxia-
induced intrauterine growth retardation. Biol Neonate 1984; 46:
103
36 Heath D, Williams DR. Physical and climatic factors at high
altitude. In: Heath D, Williams DR, eds. High-Altitude Medicine
and Pathology, 3rd edn. Oxford: Oxford University Press, 1995:
719
37 Imagawa S, Goldberg MA, Doweiko J, Bunn HF. Regulatory
elements of the erythropoietin gene. Blood 1991; 77: 27885
38 Ward MP, Milledge JS, West JB. Haematology. In: Ward MP,
Milledge JS, West JB, eds. High Altitude Medicine and Physiology.
2nd edn. London: Chapman & Hall Medical, 1995: 15572
39 Ali KZ, Burton GJ, Morad N, Ali ME. Does hypercapillarization
influence the branching pattern of terminal villi in the human
placenta at high altitude? Placenta 1996; 17: 67782
40 Mayhew TM, Jackson MR, Haas JD. Oxygen diffusive con-
ductances of human placentae from term pregnancies at low and
high altitudes. Placenta 1990; 11: 493503
41 Ballew C, Haas JD. Hematologic evidence of fetal hypoxia among
newborn infants at high altitude in Bolivia. Am J Obstet Gynecol
1986; 155: 1669
42 Beall CM. Optimal birthweights in Peruvian populations at high
and low altitudes. Am J Phys Anthropol 1981; 56: 20916
43 Haas JD, Balcazar H, Caulfield L. Variation in early neonatal
mortality for different types of fetal growth retardation. Am J
Phys Anthropol 1987; 73: 46773
44 Unger C, Weiser JK, McCullough RE, Keefer S, Moore LG.
Altitude, low birth weight, and infant mortality in Colorado.
JAMA 1988; 259: 342732
45 Moore LG, Rounds SS, Jahnigen D, Grover RF, Reeves JT. Infant
birth weight is related to maternal arterial oxygenation at high
altitude. J Appl Physiol 1982; 52: 6959
46 Moore LG, Brodeur P, Chumbe OJDB, Hofmeister S, Monge C.
Maternal hypoxic ventilatory response, ventilation, and infant
birth weight at 4300 m. J Appl Physiol 1986; 60: 14016
Ultrasound in Obstetrics and Gynecology