| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, Mount Sinai School of Medicine, One Gustave Levy Pl.,
New York, NY 10029.
2
Department of Pediatrics, Mount Sinai School of Medicine, New York, NY
10029.
3
Department of Biostatistics and Epidemiology, Cleveland Clinic Foundation,
9500 Euclid Ave., Cleveland, OH 44195.
4
Present address: Department of Epidemiology and Biostatistics, Case Western
Reserve University School of Medicine, 10900 Euclid Ave., Cleveland, OH
44106.
5
Department of Radiology, Children's Hospital, 300 Longwood Ave., Boston, MA
02115.
6
Department of Radiology, Texas Children's Hospital, 6621 Fannin St., Houston,
TX 77030.
7
Department of Pediatrics, Columbia-Presbyterian Medical Center, 630 W.168 St.,
New York, NY 10032.
8
Department of Radiology, Columbia-Presbyterian Medical Center, New York, NY
10032.
9
Department of Radiology, University of California, 10833 Le Conte Ave., Los
Angeles, CA 90095.
10
L.A.C./U.S.C., 1975 Zonal Ave., Los Angeles, CA 90033.
11
Department of Radiology, Cleveland Clinic Foundation, Cleveland, OH
44195.
12
Department of Pediatrics, Children's Hospital, 4650 Sunset Blvd., Los Angeles,
CA 90027.
Received July 20, 2000;
accepted after revision December 4, 2000.
Supported by contracts (NO1-HR-96037, 96038, 96040, 96041, 96042, and
96043) from the National Heart, Lung, and Blood Institute, and in part by NIH
General Clinical Research Center Grants (RR-00188, RR-00533, RR-00071,
RR-02172, RR-00645, RR-00865, and RR-00043).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Between 1990 and 1997, we scored 3050 chest radiographs using a standardized form. Group I children (n = 201) were HIV-1—infected at enrollment. Group II children (n = 512) were enrolled prenatally or before 28 days postpartum and subsequently subdivided into group IIa (n = 86), children identified as HIV-1-infected; and group IIb (n = 426), those who were HIV-1—uninfected. CRC were defined as parenchymal consolidations or nodular disease lasting 3 months or more or increased bronchovascular markings or reticular densities lasting 6 months or more. Morbidity was assessed by CD4 counts, viral load, the presence of low oxygen saturation, wheezing, tachypnea, crackles, and clubbing.
RESULTS. The cumulative incidence of chronic radiographic lung changes in HIV-1—infected children was 32.8% by 4 years old, with increased bronchovascular markings or reticular densities being most common. Chronic changes were associated with lower CD4 cell counts and higher viral loads. Resolution of these chronic changes was associated with decreasing CD4 cell counts but not with lower rates of clinical findings, viral load, or difference in survival.
CONCLUSION. With increased survival, CRC are becoming more common. The resolution of these changes may indicate immunologic deterioration rather than clinical improvement.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We investigated the frequency and importance of chronic radiographic lung changes (CRC) in HIV-1—infected and —uninfected children born to HIV-1—infected mothers, as part of the prospective multicenter study funded by the National Heart, Lung, and Blood Institute (National Institutes of Health), known as the Pediatric Pulmonary and Cardiovascular Complications of Vertically Transmitted HIV Infection (the P2C2 Study) [5].
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Group I consisted of 205 HIV-1 vertically infected infants and children older than 28 days and known to be infected at the time of enrollment (age range 1.7-166 months; median age at time of enrollment, 23 months). Of these, chest radiographs were obtained in 201 children. Group II consisted of 600 neonates enrolled prenatally or within 28 days of life, born to HIV-1—infected mothers. Ninety-three neonates were identified as HIV-1—infected (group IIa), of whom 86 underwent chest radiography. Four hundred sixty-three were uninfected (group IIb), of whom 426 underwent chest radiography. For the subset of 426 group IIb patients with one or more chest radiographs having been obtained, 211 were randomly selected as controls, 173 were randomized out of the study, and 42 were lost to follow-up before randomization. Forty-four had an indeterminate HIV-1 status and were excluded from further study.
There were no significant differences in gender or race between the group I and group II cohorts. More than 90% of the children in groups I and IIa took antiretroviral medication (principally zidovudine [Ritonavir; Glaxo-Wellcome, Research Triangle Park, NC] and dideoxyinosine [Videx; Bristol-Myers-Squibb, Stamford, CT]) at some time during the study. Twenty-one children in group I and seven in group IIa took protease inhibitors for up to 13 months as part of drug trials (nelfinavir [Viracept; Agouron, La Jolla, CA], saquinavir [Invirase; Roche, Basel, Switzerland], or indinavir [Crixivan; Merck, Whitehouse Station, NJ]), but all were older than 2 years when started. Forty-three percent of group I and 28% of group IIa children received IV immunoglobulin at some time during the study.
The follow-up time for comparing radiographs was calculated from enrollment to the last radiograph for group I and from birth to the last radiograph for group II. The median length of radiographic follow-up was 47.0 months for group I, 39.3 months for group IIa, and 18.1 months for group IIb. The median length of follow-up for calculating mortality was from enrollment (group I) or birth (group II) to death or loss to follow-up. The median follow-up was 54.4 months for group I, 44.1 months for group IIa, and 45.3 months for group IIb.
In group I and in group IIa, routine chest radiographs were obtained at 3, 12, and 18 months and annually thereafter. In group IIb children randomized as controls, radiographs were obtained at 3, 12, and 18 months only. Additional routine radiographs in group IIb were not obtained for ethical considerations after HIV-1 infection had definitively been ruled out. Radiographs were obtained during inter-current illnesses in all groups. The investigators, board-certified pediatric radiologists, masked to the children's HIV-1 status, evaluated 3050 chest radiographs (1989 routine and 1061 for intercurrent illnesses), using a standardized forced-choice grading form [6]. Standard radiographic terminology was used [7]. Abnormal findings on radiographs were assessed for nodular densities, reticular densities, bronchovascular markings, and parenchymal consolidations (Fig. 1).
|
Chronic lung changes on radiographs were defined as parenchymal consolidations, either focal or diffuse, that lasted 3 months or longer (Fig. 2); increased bronchovascular markings or reticular densities that lasted 6 months or longer (Fig. 3); or nodular disease that lasted 3 months or longer (Fig. 4A,4B). We chose 6 months as the cutoff for designating increased bronchovascular or reticular markings as chronic to eliminate transient findings from residual inflammatory, reactive, or obstructive pulmonary disease. Three months was chosen for consolidations and nodules by a consensus of board-certified pediatric pulmonologists and radiologists in the study. Transient radiographic changes were considered as findings that did not meet the time criteria of CRC. When one or more subsequent radiographs showed no sign of a previously identified chronic radiographic lung change, the CRC were considered resolved. The period of the CRC was called the CRC window, and one child could have several CRC windows.
|
|
|
|
Clinical examinations were performed every 3 months. Diagnoses of intercurrent illness were made as previously defined in the study design [5]. Lymphocytic interstitial pneumonia was diagnosed by biopsy. Presumptive lymphocytic interstitial pneumonia was diagnosed on the basis of clinical presentation and radiographic nodules that persisted for 2 or more months. Serum was obtained every 6 months for CD4 and viral analysis [8]. The P2C2 Mortality Review Committee determined underlying causes of death [9].
The cumulative incidence of CRC was obtained from Kaplan-Meier analyses for HIV-1—infected children (group IIa) and HIV-1—uninfected children (group IIb). In addition, the frequency of CRC for each group was reported as a proportion. Cumulative rates of CRC over time between HIV-1—infected and HIV-1—uninfected children were compared with the log-rank test.
When comparing means, we used longitudinal linear mixed-effects models [10]. In these models, a random intercept term was introduced to allow modeling of the within-subject correlation that may arise because of the longitudinal nature of the data. When comparing data in the CRC groups, we used paired t tests. Fisher's exact test or the McNemar test was used to compare proportions (such as frequency of diagnosis).
Fisher's exact test was used to compare group I and group IIa subjects with and without CRC with group IIb subjects. Radiographic findings were tested for association with clinical information, which included percentage of oxygen saturation obtained within 1 week of the radiograph, and with CD4 counts and viral load obtained within 3 months of the radiograph. Cube-root transformations for CD4 counts were made to satisfy normality assumptions [11].
Mortality for each group was compared by fitting Kaplan-Meier survival curves and then using log-rank statistics to compare the survival curves. Survival was analyzed by time since enrollment rather than by age because patients entered the study at different ages. When group IIa mortality was analyzed separately, it was analyzed by age because all group IIa patients were enrolled before birth or up to 28 days old. When additional covariate adjustments were needed for differences in the ages of enrollment of the two groups, Cox proportional hazards regression models were fit. Cumulative survival curves were constructed to compare group I and IIa subjects with CRC against group IIb subjects without CRC and against groups I and IIa children without CRC. Similar survival estimates were calculated for the HIV-1—infected children with parenchymal consolidations, bronchovascular markings or reticular densities, and nodular densities. Overall, Kaplan-Meier plots and Cox models revealed similar patterns. The frequencies of different causes of death among the HIV-1—infected children with and without CRC were also compared.
All hypothesis testing was done at a two-tailed significance level of 0.05.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
= 0.57 for parenchymal consolidations, = 0.71 for
increased bronchovascular or reticular densities, and = 0.56 for the
presence of nodular densities), and hence, the results of this study can be
generalized [6].
Bronchovascular or reticular densities were grouped together because when they
were rated separately, interreviewer reliability among radiologists was
moderate to poor [6]. In group
I, two radiographs were obtained in 15 children (7.5%), whereas three or more
radiographs were obtained in 173 children (86%). In group IIa, two radiographs
were obtained in one child (1.2%), whereas three or more radiographs were
obtained in 79 children (91.8%). In group IIb, two radiographs were obtained
in 124 children (29.1%), whereas three or more radiographs were obtained in
178 children (41.8%). Using all available gestational-age data, we determined the rate of prematurity (gestational age less than 37 weeks) to be 63 (23.8%) of 265 infants in the HIV-infected children versus 69 (16.8%) of 411 infants in the uninfected group (p =0.03). There was, however, no difference in the rate of neonatal respiratory distress in the HIV-1—infected children versus the HIV-1—uninfected children (13 [4.8%] of 273 versus 13 [3.2%] of 412, p = 0.28) or bronchopulmonary dysplasia (2 [0.7%] of 275 versus 4 [1.0%] of 410, p = 0.73).
Normal findings on radiographs throughout the entire study were present in 56 (27.9%) of 201 patients in group I, 29 (33.7%) of 86 patients in group IIa, and 379 (89.0%) of 426 patients in group IIb. The cumulative incidence of CRC in the infected (group IIa) and uninfected (group IIb) children followed up prospectively from birth is shown in Table 1. In group IIa, the cumulative incidence of CRC was 14.6% ± 4.1% at 18 months and 32.8% ± 7.1% at 48 months, with a median age of onset of 18 months. CRC were rare in group IIb. In group I, the cumulative incidence of CRC was 25.6% ± 3.2% and 36.5% ± 3.8% at 18 and 48 months since enrollment, respectively. The median age at first CRC was 35.8 months for group I children.
|
The frequency of transient and chronic lung changes for groups I and IIa is shown in Table 2. CRC were present in 31.8% of group I. In the prospective cohort followed up from birth, group IIa, 22.1% had CRC.
|
Bronchovascular markings or reticular densities were the most frequently reported finding in all groups. The frequency of each type of CRC was higher in the group I children compared with the group IIa children. The proportion of each chronic abnormality was similar for groups I and IIa. In groups I and IIa, respectively, parenchymal consolidations were seen in 24 (37.5%) of 64 versus seven (36.8%) of 19 of children with CRC; bronchovascular markings or reticular densities were seen in 40 (62.5%) of 64 versus 12 (63.2%) of 19; and nodular densities were seen in 26 (40.6%) of 64 versus seven (36.8%) of 19. More than one chronic abnormality was seen in 32% of children.
Because the proportion of each type of CRC was similar in group I and group IIa children, these groups were combined for analysis of the relationship of CRC to the clinical findings. Significant increases (p < 0.001) in frequency of clubbing, crackles, and low levels of oxygen saturation were found in HIV-1—infected children with CRC versus HIV-1—uninfected children. No statistically significant increase in wheezing frequency was observed for any type of CRC. Low levels of oxygen saturation were more common in children with parenchymal consolidations (20 [64.5%] of 31). Crackles were more likely to occur during a parenchymal consolidations window (14 [48.3%] of 29) than during any other CRC window, and tachypnea was most likely to occur during a nodular densities window (16 [48.5%] of 33).
During the follow-up period, CRC resolved in 58% of the HIV-1—infected children with parenchymal consolidations, in 50% of those with bronchovascular markings or reticular densities, and in 61% of those with nodular densities. The resolution of CRC was not associated with significant improvement in oxygen saturation or resolution of clubbing, crackles, or tachypnea (all p values with the McNemar test were p > 0.05).
The HIV-1—infected children with CRC had significantly lower mean CD4 counts than HIV-1—infected children without CRC (Fig. 5). CD4 counts were significantly lower after the resolution of CRC than they had been before and during the CRC window. When all types of CRC were grouped together, CRC resolution was associated with a mean drop in CD4 counts of 146 ± 62 cells/mm3 (mean drop ± SEM) (p = 0.01). Similar results were found when each type of CRC was analyzed individually.
|
The viral load in HIV-1—infected children with CRC was higher than the viral load in HIV-1—infected children without CRC. However, the resolution of CRC was not associated with any significant difference in viral load (p = 0.29).
Pneumocystis carinii pneumonia was diagnosed in five (6.0%) of 83 HIV-1—infected patients with CRC. All patients were children with parenchymal consolidations, and these five children represented 16.1% of the 31 patients with parenchymal consolidations. In the HIV-1—infected CRC subgroup, lymphocytic interstitial pneumonia was documented in four (4.8%), and presumptive lymphocytic interstitial pneumonia in 12 (14.5%) of 83. All children with presumptive lymphocytic interstitial pneumonia had nodular densities. Three of four with biopsy-proven lymphocytic interstitial pneumonia had nodular densities, and one of four had parenchymal consolidations. No child in the HIV-1—uninfected group was diagnosed with lymphocytic interstitial pneumonia or presumptive lymphocytic interstitial pneumonia (both p < 0.001). Tuberculosis was diagnosed in only two (2.4%) of 83 of the HIV-1—infected group with CRC (one child with parenchymal consolidations and one child with nodular densities).
Mortality among HIV-1—infected children, regardless of CRC, was significantly higher than among HIV-1—uninfected children (one recorded death). Of the 287 HIV-1—infected children, 95 died: 31 with CRC and 64 without CRC. Mortality among the HIV-1—infected children with CRC was not significantly different from mortality in HIV-1—infected children without CRC (Fig. 6). When the HIV-1-infected children with CRC were broken down by study group, CRC status did not affect mortality (group I, p = 0.26; group IIa, p = 0.61; by log-rank test). There was no change in mortality rate in those HIV-1-infected children with CRC whose CRC resolved.
|
Of the 31 deaths that occurred in the HIV-1-infected children with CRC, only 10 deaths (32.2%) were from pulmonary infection. Twelve (18.8%) of the 64 deaths of HIV-1-infected children without CRC were attributed to pulmonary infection. When comparing deaths from pulmonary causes between the groups with CRC and without CRC, we found no significant difference (p = 0.14 by log-rank test).
Of the 83 children in this study with CRC, 10 were administered protease inhibitors for up to 13 months as part of drug trials, but nine of the children had CRC before beginning the drug trials.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The study reflects contemporary clinical realities in two ways. First, although CT might have produced more detailed information about lung conditions [14, 15], we studied chest radiographs because they were more common, practical, and less costly than CT scans.
Second, lung changes in infants caused by viral infection might persist for months. To avoid any confusion between these transient changes and the chronic disease associated with HIV-1 progression, we agreed on conservative definitions of chronic change (persistence for 3 or 6 months). We cannot exclude the possibility that two separate intercurrent illnesses occurred during the CRC window and that this combination may have caused some overestimation of CRC. This scenario is not likely, however, because abnormal findings on radiographs were clinically followed up, as per a defined protocol for intercurrent illnesses: three or more radiographs were obtained in 86% of the group I children and in 91.8% of the group IIa children during the study.
CRC are common in HIV-1-infected children, and the cumulative incidence increases with age. The frequency of each chronic radiographic change (parenchymal consolidations, bronchovascular markings or reticular densities, and nodular densities) was higher in the group I (HIV-1-infected enrolled at any time after the age of 28 days) compared with the group IIa patients (HIV-1-infected followed up from birth), although the proportion of children in each group with each type of change was similar. Group I children had the same duration of follow up as group IIa but were older and included patients who had lung disease at the time of enrollment in the study. The median age of group I children with CRC at onset of CRC was 35.8 months compared with 18 months in group IIa. Therefore, selection bias likely accounts for the higher frequency of CRC seen in group I. Respiratory disease as a result of prematurity was excluded as a cause of CRC because the incidence and severity of major neonatal respiratory disorders was similar in the infected and uninfected groups [16].
Bronchovascular markings or reticular densities were the most common chronic changes in the HIV-1-infected children. These radiographic patterns can be seen in children with asthma. Although some children in this study with these changes may have had asthma, the changes were not associated with wheeze or tachypnea. Bronchovascular markings or reticular densities did correlate with the clinical findings of crackles, clubbing, and oxygen desaturation, which can be observed with chronic aspiration or interstitial pneumonia. The nature of these changes requires further study.
We recognize that in clinical practice, nodular changes identified on radiographs are generally not biopsied to determine whether they represent lymphocytic interstitial pneumonia or pulmonary lymphoid hyperplasia. The Centers for Disease Control and Prevention define lymphocytic interstitial pneumonia or pulmonary lymphoid hyperplasia as nodular disease persisting for more than 2 months, with no need for biopsy confirmation [17]. In group I, 32% had nodular densities. Rates as high as 40% have been cited in other nonprospective studies [12, 18]. In our prospective cohort followed up from birth, group IIa, only 8% developed nodular densities persisting for more than 3 months during the follow-up period. The higher rate of nodular densities in group I most likely represents selection bias. In the absence of lung biopsies in all cases of nodular densities, we cannot be certain if these nodular densities represent lymphocytic interstitial pneumonia or pulmonary lymphoid hyperplasia, especially in light of the fact that they resolved in some patients. Nevertheless, it is clear that the prevalence of nodular densities is low in HIV-1-infected children in the first few years of life.
The causes of chronic consolidations in children with HIV-1 infection include bronchiectasis, chronic interstitial pneumonitis, and chronic atelectasis and disordered mucociliary clearance [13]. The P2C2 data indicate that chronic parenchymal consolidations in HIV-1-infected children can be a manifestation of P. carinii pneumonia. This is consistent with a recent report in adults that noted that P. carinii pneumonia might appear as a chronic infiltrate, caused by a form of interstitial fibrosis [19].
HIV-1-infected children with CRC had lower CD4 counts and a higher viral load than HIV-1-infected children without CRC. However, we found that CRC did not affect survival within the time limits of this study. In addition, CRC resolved in many patients and were accompanied by significantly lower CD4 counts, suggesting that in HIV-1-infected patients, the ability to mount an immunologic reaction within the lungs and produce an abnormality revealed on radiography may decrease over time. This resolution was not associated with any clinical improvement in frequency of clubbing, crackles, tachypnea, or episodes of low levels of oxygen saturation. The P2C2 study data are in agreement with previous reports speculating that the resolution of nodular changes in the HIV-1-infected child may be an indicator of immunologic deterioration rather than clinical improvement [13, 15, 20]. However, we did not document that the resolution of CRC was associated with a change in viral load.
The time interval selected to correlate radiographic findings with measurements of CD4 counts and viral load was 3 months. Viral load remained relatively stable after patients were 2 months old [21]. CD4 counts decline slowly after 15 months and were nearly stable by 21 months [22]. Most of the children with CRC were older than 1 year, and thus, the mean HIV-1 RNA and mean CD4 counts had reached stable levels. Therefore, it is unlikely that large fluctuations in CD4 counts and viral load would have occurred during the 3-month time interval.
This study was conducted between 1990 and 1997, a period in which neonates were being diagnosed with HIV-1 at birth and mothers were given zidovudine in the perinatal period, resulting in a lower mother-to-infant transmission rate. Since the conclusion of the study, highly active antiretroviral therapy has become a part of treatment regimens. The introduction of protease inhibitors has altered the progression of disease in many HIV-1-infected patients. Most patients in our study were on antiretroviral therapy but not on protease inhibitors. It is possible that with newer pharmacologic regimens, the resolution of CRC may no longer represent a sign of immunologic decline, but rather a favorable prognostic finding.
In the early years of the AIDS epidemic, mortality from acute lung disease, particularly P. carinii pneumonia, was high [23]. Today, early diagnosis, P. carinii pneumonia prophylaxis, and better treatment have lowered mortality rates from acute lung disease but have increased the number of HIV-1-infected children with CRC. Determining how these chronic changes correlate with morbidity or mortality might lead to new screening strategies or therapeutic modalities. It is probable that with longer survival, CRC will supplant opportunistic pulmonary disease as the major cause of pulmonary complications in the HIV-1-infected child.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
3 months, bronchovascular markings or
reticular densities