Vertical transmission of HIV from HIV-positive mothers to their infants is, in industrialised countries, reduced to less than 1-2% [
The thymus plays a key role in the development of a functional immune system, providing the environment for T-lymphocyte maturation and being a central organ for the development and maintenance of cell-mediated immunity. The thymus is also known to be a target organ in HIV-infection [
The positive correlation between thymic size and weight and length at birth and during the first months of life is well known and infections are reported to result in decreasing thymic size [
In HIV-infected children, thymic size is smaller compared to HIV-negative children, as reported in two African studies [
Few studies have managed to demonstrate an association between thymic size and T-lymphocytes. In the two African studies of HIV-infected children, the thymic volume was found to correlate positively with the CD4 count and negatively with HIV viral load [
The present study was conducted to evaluate the thymic size in HIV-EU infants compared to term HIV-unexposed infants. Furthermore, the impact of prenatal HIV exposure on HIV-negative infants fed with human donor milk in their first four months of life on the thymic size and frequency of infections during the first year of life was evaluated. Finally, we aimed to examine a possible correlation between ultrasound-obtained thymic size, T-lymphocyte subsets in peripheral blood, and the frequency of infections.
From December 1996 to October 1999, all infants born to HIV-positive mothers at Hvidovre University Hospital were invited to participate in the study. One mother declined to take part in the study. A cohort of 18 infants born to HIV-positive mothers was included in the study. The study population included one pair of twins. All but one infant were delivered by Caesarean section. The 17 HIV-positive mothers included 2 ethnic Danes (including 1 former intravenous drug user (IVDU)), 12 women from Africa, and 3 women from Asia. Antiretroviral treatment antepartum and/or intrapartum was accepted by all the HIV-positive mothers. None of the infants were infected with HIV as detected by PCR at the age of 6 months. One infant born to an HIV-positive mother was born preterm at a gestational age (GA) of 31 weeks (GA being the age of the infant determined as the number of weeks after the first day of the last menstrual period). The infants were examined at birth and after 2 (
All HIV-EU infants were exclusively fed pasteurized human milk (donor milk) provided from the local milk bank: The Hvidovre Milk Bank, which is the largest milk bank in Denmark. The milk is obtained from mothers who are referred to the milk bank by health visitors. All donor milk mothers were initially interviewed at the milk bank and all were screened for HIV, hepatitis B, and hepatitis C. Milk from immigrants was screened for TB. The donor milk mothers come from all over the island of Sjaelland, include all ethnicities, and were paid per liter for their milk. All the milk was pasteurized and controlled according to protocols and guidelines that were part of the public health law. For further information please see the article by Arnold in 1999 [
The thymic size and clinical outcome in the HIV-EU infants were compared to a cohort of 47 term HIV-unexposed infants from an earlier thymus cohort study conducted in the period October 1994 to December 1996 [
Blood samples were collected and used for analysis of lymphocyte subsets by flow cytometry. A total leukocyte count and a differential white cell count were obtained to calculate the absolute numbers of lymphocyte subsets. Flow cytometry was performed as described previously [
Lymphocyte subsets were not analysed in the control cohort.
The size of the thymus was assessed by ultrasound as the thymic index (Ti) as previously described by Hasselbalch et al. [
Descriptive analysis was performed by means of mean and median values and range (i.e., mix–max) for Ti, weight, length, and cumulated number of infections and T-lymphocytes.
The correlation between Ti, weight, length, and cumulated number of infections and T-lymphocytes was calculated using Spearman’s correlation coefficient.
Differences in Ti at birth and at 12 months of age between the HIV-EU and the HIV-unexposed infants were evaluated using an analysis of variance. Ti was the outcome variable and exposure (yes, no) was included as a fixed effect. Due to a skewed distribution of Ti, a logarithmic transformation of Ti values was used in the statistical analysis. The assumptions were evaluated using the Shapiro-Wilks test for a normal distribution and visual inspection of residual plots.
A repeated measurements analysis of variance was performed to evaluate differences in Ti between feeding from 0 to 12 months of age. Ti was the outcome variable. Feeding group and month were included as fixed effects. Differences in the development of Ti between feeding groups during the study period were evaluated by including the interaction between month and feeding group. The autocorrelation between repeated Ti values for the same infant was taken into account by including an autoregressive correlation structure in the model. The assumptions were evaluated using the Shapiro-Wilks test for a normal distribution and visual inspection of residual plots.
The association between the cumulated number of infections and Ti was evaluated in a repeated measurements analysis with cumulated number of infections as the outcome variable. A Poisson regression model was used. Ti, month, and feeding group were included as fixed effects. The autocorrelation between repeated numbers of infections for the same infant was taken into account by including an autoregressive correlation structure in the model. The dispersion parameter was used to evaluate the goodness-of-fit for the model.
The association between T-lymphocyte subsets and Ti was analyzed using a repeated measurements analysis of variance with each of the T-lymphocyte subgroups as the outcome variable. Ti and month were included as fixed effects. The autocorrelation between repeated T-lymphocyte values for the same infant was taken into account by including an autoregressive correlation structure in the model. The assumptions were evaluated using the Shapiro-Wilks test for a normal distribution and visual inspection of residual plots.
A 5% significance level was used in all analyses. All statistical analyses were performed using SAS Statistical Analysis Software (version 9.2).
Ti in HIV-EU infants and in the cohort of term HIV-unexposed infants is shown in Tables
HIV-exposed un-infected infants (HIV-EU): the number of infants, the thymic size (thymic index (Ti)), and the weight and length at the times of assessment.
Time of examination | Newborn | 2 months | 4 months | 8 months | 12 months | 18 months |
---|---|---|---|---|---|---|
Number | 18 | 13 | 13 | 10 | 9 | 6 |
Ti (median) | 9.3 |
17.2 |
24.0 |
24.3 |
20.0 |
21.8 |
Weight, (g) mean |
2926 |
4159 |
6005 |
8216 |
9311 |
10660 |
Length, (cm) mean |
49.7 |
54.7 |
60.7 |
68.2 |
74.2 |
77.9 |
The term HIV-unexposed infants: the number of infants, the thymic size (Ti), and the weight and length at the times of assessment.
Time of examination | Newborn | 2 months | 4 months | 8 months | 12 months |
---|---|---|---|---|---|
Number | 47 | Not examined | 47 | 37 | 37 |
Ti (median) | 11.8 |
Not examined | 27.7 |
29.0 |
17.3 |
Weight, (g) mean |
3524 |
Not examined | 7091 |
9222 |
10442 |
Length, (cm) mean |
52.0 |
Not examined | 64.7 |
72.3 |
76.9 |
The thymic size in the HIV-EU infants at birth tended to be lower than thymic size in the term infants at birth, although this did not reach significance (
The thymic size in HIV-EU infants compared to term HIV-unexposed infants. In general, there is no difference in thymic size between the cohorts from birth to 12 months of age when feeding mode is not included in the analysis. The Ti is given as median and with Q1 and Q3.
There was a significant difference in Ti between the four feeding groups across the examination period 0–12 months of age (
The term HIV-unexposed infants divided into the three feeding groups at 4 months of age: the number of infants, the thymic size (Ti, median, and range), and the weight and length (mean and range) at the times of assessment.
Three feeding groups | Values | Newborn | 4 months | 8 months | 12 months |
---|---|---|---|---|---|
Exclusively breastfed | Number | 21 | 21 | 15 | 15 |
Exclusively breastfed | Ti | 12.9 |
37.8 |
30.6 |
17.3 |
Exclusively breastfed | Weight (g) | 3446 |
7237 |
9262 |
10375 |
Exclusively breastfed | Length (cm) | 52.1 |
64.9 |
72.3 |
77.7 |
Partially breastfed | Number | 13 | 13 | 11 | 11 |
Partially breastfed | Ti | 11.2 |
25.2 |
25.6 |
17.4 |
Partially breastfed | Weight (g) | 3658 |
6917 |
8887 |
10114 |
Partially breastfed | Length (cm) | 52.7 |
64.3 |
71.3 |
75.4 |
Exclusively formula-fed | Number | 13 | 13 | 11 | 11 |
Exclusively formula-fed | Ti | 9.7 |
17.6 |
30.4 |
17.0 |
Exclusively formula-fed | Weight (g) | 3518 |
7029 |
9503 |
10861 |
Exclusively formula-fed | Length (cm) | 52.4 |
64.7 |
73.2 |
77.5 |
Feeding mode at 4 months of age had a significant impact on the thymic size. Those breastfed at 4 months of age have the largest thymus when compared to the three other feeding groups. However, the HIV-EU infants at 4 months of age have a thymic size the same as the partially breastfed infants. By 8- and 12-month followups the difference in Ti between the feeding groups disappeared. Ti is given as medium and with Q1 and Q3.
There was no overall correlation between Ti and number of infections at 0, 4, 8, 12, and 18 months of age (
There was a significant difference in the cumulated number of infectious events between the feeding groups across the examination period 0–12 months of age (
Events of infectious illness in HIV-exposed uninfected (HIV-EU) infants compared to term HIV-unexposed infants, according to their different feeding modes at 4 months of age. The observation period between each examination was 4 months.
Mean number (min–max) of cumulated infectious events | |||
---|---|---|---|
Time of examination: |
4 months | 8 months | 12 months |
Feeding mode and cohort: | |||
HIV-EU: human donor milk | 1.8 (0–8) | 1.7 (0–4)* | 3.2 (0–7)* |
HIV-unexposed infants: exclusively breastfed until 4 months of age | 2.0 (0–4) | 5.7 (1–13)** | 9.0 (5–20) |
HIV-unexposed infants: exclusively formula-fed until 4 months of age | 2.5 (1–5) | 12.5 (2–35) | 9.5 (3–15) |
HIV-unexposed infants: partially breastfed until 4 months of age | 1.8 (1–3) | 8.2 (1–26) | 7.9 (4–13) |
Among the HIV-unexposed infants, the exclusively breastfed infants had significantly fewer infectious events at 8 months of age than the formula-fed infants at the same age (
Table
Distribution of T-lymphocyte subsets in HIV exposed uninfected infants given as median and range (min–max) and the number of infants (
T-lymphocytes | Newborn |
2 months |
4 months |
8 months |
12 months |
18 months |
---|---|---|---|---|---|---|
CD3+ | 2575 |
3340 (2094–5281) | 3340 (1995–4697) | 3451 (2548–5059) | 3433 (2043–5396) | 3825 (1686–5073) |
CD4+ | 1109 (325–2157) | 1370 (803–2855) | 1468 (83–2049) | 1426 (962–1967) | 1325 (849–2502) | 1263 (658–2488) |
CD8+ | 463 (65–2895) | 642 (306–2991) | 467 (128–1264) | 634 (263–1708) | 631 (242–1423) | 551 (254–902) |
CD1a+ CD7+ | 9.1 (3.6–65.8) | 26.2 (3.8–126.2) | 26.5 (10.4–114.6) | 25.3 (6.1–83.5) | 30.1 (7.0–98.9) | 17.7 (8.3–69.0) |
A high number of CD3+, CD8+, and CD4+ cells was significantly associated with fewer infectious events during the observation period (
The purpose of this study was to evaluate the thymic size and infections in HIV-EU human donor milk fed infants compared to term HIV-unexposed infants. In general, equal growth and size of thymus was found in the HIV-EU infants compared to term HIV-unexposed infants. Interestingly, HIV-EU infants had fewer infections up to one year of life, when compared to the HIV-unexposed infants. Finally, in the HIV-unexposed infants we demonstrated fewer infections when exclusively breastfed at 4 months of age.
This birth cohort of HIV-EU infants was unique in Danish settings, being followed prospectively and given human donor bank milk for their first 4 living months. The possibility of giving donor milk to HIV-EU infants had ceased in 2003, because of high economic cost.
The thymic size was evaluated by ultrasound. The method was first described by Hasselbalch et al., and although thymic index is a volume estimate and not a true volume, it was shown by a postmortem study to correlate well with actual thymic volume [
At birth, the thymic size tended to be smaller in the cohort of HIV-EU infants (when the preterm infant was included) compared to the cohort of HIV-unexposed infants. However, when feeding mode was not taken into account the thymic size at one year of age in this cohort of HIV-EU infants did not differ significantly from the whole cohort of term healthy HIV-unexposed infants at the same age.
This is in contrast to a recent study of HIV-EU infants examined at around 15 months of age, who had a significantly smaller thymic size compared to the matched cohort of infants born to HIV-negative mothers [
These conflicting findings might be explained by the different feeding modes of the two cohorts of HIV-EU infants; the infants in this presented study received human donor milk in the first 4 months of life, compared to the other cohort of HIV-EU infants who were exclusively formula-fed from birth [
Interestingly, HIV-EU infants had significantly fewer infections up to one year of life when compared to term HIV-unexposed infants. This finding may be explained by the feeding mode too, but even compared to the exclusively breastfed infants the rate of infections was lower. Another potential explanation could be differences in attending day-care or number of older siblings between the two cohorts. Unfortunately, we do not have data on day-care attendance or number of siblings. Also, the lower the number of infections in the HIV-EU infants could be explained by different socio-economic status, but the mothers from both cohorts came from the same suburbs of Copenhagen. Finally, the ethnical difference between the HIV-EU and the control infants might influence both occurrences of infections and thymic growth, but to our knowledge no Danish studies have shown difference in occurrence of infections between different ethnic groups of infants in Denmark.
It could be discussed whether parental recall for infections is reliable [
One might speculate that being exposed to HIV-virus in utero might prime the immune system in the same way as vaccines are suggested to have immune modulation properties influencing child mortality in African infants [
The correlation between CD4+ and CD8+ cells and thymic size demonstrated in previous studies was not observed in this study [
This study shows that HIV-EU infants fed human donor milk are capable of a normal growth of thymus and contract fewer infections than other infants. Furthermore, the finding of fewer infections at 8 months of age in the HIV-unexposed infants who were exclusively breastfed to 4 months supports evidence for a beneficial effect of human milk on the immune system. We suggest that when breastfeeding is not possible, that providing vulnerable groups of infants human donor milk will be beneficial for the maturing immune system.
This study was approved by the local Scientific Ethics Committee for the Municipalities of Copenhagen and Frederiksberg, Denmark (KF 01-165/96), and informed consent was obtained from all patients after the nature and consequences of the study had been fully explained.
There is no conflict of interests to declare.
The authors would like to thank the participants and their families for their participation in the study. The project was financially supported by Grants from the Foundations of Sygekassernes Helsefond, Det Sundhedsvidenskabelige Forskningsråd, and Fonden af 17-12-1981.