Learn more: PMC Disclaimer | PMC Copyright Notice
Zinc Modifies the Association between Nasopharyngeal Streptococcus pneumoniae Carriage and Risk of Acute Lower Respiratory Infection among Young Children in Rural Nepal1,2
Abstract
Nasopharyngeal (NP) carriage is necessary for Streptococcus pneumoniae (Spn) transmission and invasive infection. This study evaluated the effect of zinc prophylaxis on the association between NP colonization with Spn and acute lower respiratory infection (ALRI) in children aged 1–35 mo living in a rural district in southern Nepal. We compared carriage prevalence of Spn in 550 ALRI cases with that of healthy age- and season-matched controls. This study, conducted from December 2003 to July 2005, was nested in a community-randomized trial designed to evaluate the effect of zinc on morbidity and mortality in 1- to 36-mo-old children. They were randomized to receive either 10-mg tablets of zinc or placebo daily until discharge. Approximately 75% of cases and controls were Spn carriers. There was an interaction between zinc and Spn carriage (P = 0.091). Spn carriage increased the risk of ALRI in the placebo group [adjusted matched odds ratio (AMOR) = 2.57; P = 0.025] but not in the zinc group (AMOR = 0.95; P = 0.890). Among the subset of symptomatic cases and their controls, the odds of ALRI for Spn carriers in the placebo group was 30 times greater (AMOR = 78.09; P = 0.006) than in the zinc group (AMOR = 2.77; P = 0.288). These findings suggest that zinc prophylaxis may protect children against ALRI associated with carriage of Spn and that the effect may differ by infectious etiology.
Introduction
Globally, pneumonia remains a leading cause of morbidity and mortality in young children, especially among those in developing countries (1,2). Although there are no precise figures, Streptococcus pneumoniae (Spn)6 is considered to be among the most common causes of severe pneumonia in this risk group (2,3).
Asymptomatic nasopharyngeal (NP) carriage of Spn, which is prevalent in early infancy in developing countries, is the first step in the infection cycle (4,5). Spn conjugate vaccines are highly efficacious for the prevention of invasive disease caused by vaccine strains (6,7). Yet their greatest impact is in lowering disease transmission by reducing NP carriage of these strains (7,8). However, most developing countries have not introduced these vaccines, because they are expensive and there are concerns over the regional relevance of strains included in current conjugate vaccine formulations. In light of these issues, research on affordable and effective community-based pneumonia prevention strategies, which can be used alone or in combination with vaccination, is needed to reduce the risk of Spn pneumonia and its transmission (9).
There is growing evidence of zinc's critical role in the regulation and maintenance of host defense systems that protect against pneumonia and other infections. Zinc deficiency is common in children in South Asia because of low dietary intake, limited bioavailability from local diets, and losses during infection. Previous studies indicated that zinc-deficient children have a higher incidence of infections, including acute lower respiratory infection (ALRI) (10,11). Results from recent prophylaxis trials that have evaluated the impact of zinc on ALRI incidence among young children in South Asia have shown variation in treatment effect. Three trials have reported significant overall protective effects (12–14), ranging from 45 to 15% reduction in ALRI risk. Another study found a protective effect only in children with poor zinc status at baseline (15). Authors of a pooled analysis of the 4 prophylaxis trials concluded that zinc reduces the incidence of clinical pneumonia by 20% (16). A large, recently completed, community-based trial in Nepal, however, found no evidence that ALRI risk differed between zinc and placebo groups (17).
Whether the protective effect of zinc in ALRI can be partially attributed to reducing NP colonization with pathogenic bacteria or to altering the relationship between carriage and infection is not yet known. Data from previous studies suggest that zinc supplementation helps restore local and systemic antibody-mediated responses in zinc-deficient animals (18–20). If zinc is found to block colonization or interfere with the infection process, then it may be an affordable approach for preventing severe ALRI. In addition, differences in the etiology of ALRI by population may help to explain differences in effect size between trials. We conducted a population-based, prospective, matched case-control study to evaluate the effect of zinc prophylaxis on the association between NP carriage with Spn and the risk of ALRI.
Materials and Methods
Study population
The matched case-control study was nested in the large, cluster-randomized, double-blind, placebo-controlled, 2 × 2 factorial, community-based Nepal Nutrition Intervention Project, Sarlahi-4 (NNIPS4) trial. The trial was conducted in southern Nepal between October 2001 and January 2006. The objectives of NNIPS4 were to evaluate the efficacy of zinc, iron, and folic acid prophylaxis on morbidity, mortality, and growth in 41,276 children aged 1–36 mo in the rural district of Sarlahi. Children living in households within 30 village development committees (VDC) in Sarlahi were cluster-randomized to receive daily doses of either zinc (10 mg), iron (12.5 mg) and folic acid (50 μg), or placebo. Children < 12 mo of age received one-half the prescribed dosage. This area was selected because of endemic micronutrient deficiency, a high childhood incidence of infections, and demographic similarities with other poor, rural South Asian communities. In November 2003, prior to the initiation of the case-control study, the iron/folic acid arms of the trial were stopped on the recommendation of the trial's data and safety monitoring board due to the absence of any treatment effect on study outcomes. Sectors that had been randomized to iron/folic acid were then rerandomized to zinc or placebo and consent was again obtained from the parents of children in these areas. Data from the trial suggest that the magnitude of zinc deficiency in the study population was high; 42% of children in the placebo group had serum zinc concentrations < 10.7 μmol/L. The serum zinc concentration was 11.0 ±2.1 μmol/L in the placebo group and 11.8 ±2.4 μmol/L in the zinc group 12 mo after supplementation was initiated. ALRI-associated morbidity and mortality did not differ between children in the 2 groups. The findings of the NNIPS4 trial have been published (17,21).
NNIPS4 enrollment
All 1- to 36-mo-old children living in households within the study area were eligible to participate in the NNIPS4 trial. Enrollment was strictly voluntary. Children were enrolled upon receipt of informed consent from a parent or guardian. Following enrollment, interviewers recorded data on household characteristics, socioeconomic indicators, and mid-upper arm circumference of the study children prior to the initiation of supplementation. Staff members were responsible for visiting the children at home twice weekly to give their assigned daily supplements. A sufficient supply of tablets was left with caregivers, who were responsible for supplementing children on days between household visits by study field staff. Annually, ∼1200 children < 24 mo of age were randomly selected for participation in a morbidity substudy. Field workers visited substudy participants every week for 12 mo to record information on the presence, onset, and duration of signs and symptoms of specific morbidities, including ALRI. All trial participants were followed until death, out-migration from the study area, or discharge at age 36 mo.
Identification and ascertainment of ALRI cases for PNP study
The PNP study was conducted from December 2003 to July 2004. All children <36 mo of age participating in the morbidity substudy of the NNIPS4 trial during those 2 periods and who met the criteria for ALRI were eligible for enrollment as cases. A case of ALRI was defined by presence of all 3 criteria: 1) cough; 2) fever or shaking chills; and 3) fast breathing, difficulty breathing, or chest retractions. There were no a priori exclusions based on family or child characteristics. Thus, the cases represent the general population of young children in Sarlahi with ALRI.
Cases were identified prospectively. During their weekly household morbidity visits, field workers elicited and recorded information from parents regarding their children's illness history for the previous 7 d. Parents of children who fulfilled the ALRI case criteria on ≥1 of the 7 recalled days were given information about the carriage study. Following the morbidity visit, the interviewer provided details about potential cases to trained specimen collectors, who returned to conduct a PNP enrollment visit. The collectors explained the case-control study to parents and requested the participation of their children. When a case was enrolled, collectors recorded current signs and symptoms of ALRI, noted whether or the not the child had received treatment for any illness in the prior 1–7 d, and collected an NP sample from the child. All children with symptomatic illness were referred to the nearest health center for treatment. Once a child had been enrolled as a case, at least 4 wk had to elapse since they last met the ALRI criteria to be enrolled as a new case.
Selection of controls
After obtaining a NP sample from a case, collectors identified potential controls from a list of substudy participants from the same VDC. Controls comprised children aged 1–36 mo who had not been an ALRI case in the previous 4 wk and were living in the morbidity substudy areas. The controls were individually matched to cases on age (±3 mo), season, and VDC. If there were no appropriate age- and season-matched controls within the same VDC as the case, a suitable control was selected from the nearest VDC. The collectors attempted to recruit controls within 7 d of enrolling the cases. Once enrolled, the collectors assessed the control for symptoms of ALRI and collected a NP sample. If a selected control met the ALRI criteria during the specimen collection visit, the collector enrolled the child as a new case and identified new controls using the procedures described above.
Ethical review
The NNIPS4 trial and case-control study protocols were approved by the Nepal Health Research Council and the Committee for Human Research of the Johns Hopkins Bloomberg School of Public Health, Baltimore, MD. Children were included in each study based on oral informed consent from parents or guardians. Verbal consent was considered appropriate given the low literacy level in the community.
Data collection
NP specimen collection.
Specimens were obtained following a set protocol. A small, flexible, rayon-tipped swab (Fisherbrand Calcium Alginate Swabs, Fisher Scientific) was inserted into 1 of the nares to the level of the posterior nasopharynx. The swab was left in place for 5 s or rotated 180 degrees before removal. Swabs were then inserted into skim milk media and the specimens were transported on ice to the field office. The specimens were stored at −20°C in a freezer at the Sarlahi field office and shipped frozen to the microbiology laboratory at the Institute of Medicine in Kathmandu on a weekly basis.
Laboratory procedures
For the isolation of Spn, swabs were inoculated onto blood agar (Becton Dickinson) plates containing 5% sheep blood and 2.5 mg/L gentamicin (Nathan Pirumal). Plates were incubated at 37°C in 5% CO2 for 18–24 h. Colonies exhibiting classic Spn morphology were confirmed by optochin (Taxo) inhibition or bile solubility testing. Spn reference strains (American Type Culture Collection 5603) were used for quality control. Isolates were serogrouped/typed using PNEUMOTEST kits (Staten Seruminstitut). The antisera included in the kits reacts with serotypes 1, 2, 3, 4, 5, 8, 14, and 20 and with serogroups 6, 7, 9, 11, 12, 15, 17, 18, 19, 22, 23, and 33. Isolates that could not be serotyped using the kits underwent confirmatory serotyping at the Staten Seruminstitut in Denmark. Serotypes were classified based on their inclusion in the currently licensed 7-valent Spn conjugate vaccine (Wyeth Vaccines) or 2 Spn conjugate vaccines under evaluation for use in low-income countries: a 10-valent (GlaxoSmithKline) and a 13-valent (Wyeth Vaccines) vaccine. The 7-valent vaccine includes serotypes 4, 6B, 9V, 14, 18C, 19F, and 23 F. The 10-valent vaccine includes all of the previous 7 along with serotypes 1, 5, and 7F. The 13-valent vaccine includes all 10 serotypes with the addition of serotypes 3, 6A, and 19A.
Statistical analyses
The incidence of ALRI was assumed to be 1.2/child-year, resulting in ∼1080 ALRI cases per year among the ∼1800 children who were 1–36 mo old in the morbidity substudy area during the follow-up period. Our goal was to enroll at least 440 ALRI cases and 440 matched controls for a total of 880 children. The sample size had the power to detect an odds ratio of 1.53 to test the null hypothesis that Spn carriage is not associated with ALRI (α level = 0.05, power = 0.80, estimated coefficient of exposure between cases and controls = 0.20, and an estimated Spn carriage prevalence in controls = 50%).
The association between bacterial carriage and case status was stratified by treatment group and analyzed on an intention-to-treat basis. Data were analyzed using Stata 9.0 (Stata Corporation). Two-tailed McNemar's tests and paired Student's t test were used, as appropriate, to compare baseline characteristics and exposures between cases and age- and season-matched controls. We used conditional logistic regression models, stratified by treatment group, at the bivariate level to evaluate potential risk factors, covariates, and confounders. Spn carriage status variables were entered into multivariate conditional logistic regression models stratified by treatment group and adjusted for the influence of covariates and confounders. Odds ratios and 95% CI were used to measure the association between exposures and disease status. P ≤ 0.05 was considered significant in the multivariate analyses. Interaction terms were significant at P ≤ 0.1.
Results
Enrollment.
NP samples were collected from 604 ALRI cases and 604 age- and season-matched controls enrolled in the carriage study. Of this number, NP swabs from 54 (8.9%) cases were excluded from the analysis, because the NP specimens had been collected >7 d after the last report of ALRI symptoms. Thus, a total of 550 cases and 550 matched controls are included in the analysis of Spn carriage.
Participant characteristics.
Approximately 50% of cases were males aged (mean ±SD) 17.2 ±7.4 mo. Nearly all were Hindu, >85% were from the Madeshi ethnic group (i.e. originating from the plains region of Nepal), and ∼80% were from low-caste families. A similar proportion of the mothers of cases had not received any schooling and <16% of mothers reported a history of smoking. Nearly 80% of cases had 1 or more siblings. Few households had electricity (17.1%). Whereas most families owned a small parcel of land (70%), ∼40% indicated that they were of very low socioeconomic status. Enrollment of cases tended to follow the seasonal pattern of lower respiratory illness, with the highest proportion being enrolled in the cool (41.6%) and hot (48.6%) seasons. Controls were comparable to cases with regard to demographic characteristics, socioeconomic indicators, and season of enrollment. However, there were a few notable differences: controls (10.2%) were less likely than cases (15.6%) to have a mother who smoked (P = 0.015) and also less likely to have 1 or more siblings <5 y old (75.2 vs. 79.8%, respectively; P = 0.055) compared with cases. In addition, the prevalence of Spn carriage was lower in ALRI cases (76.3% ) than in controls (81.8%; P = 0.035).
Stratified bivariate analyses.
We compared baseline characteristics, Spn carriage status, and whether or not the child had received treatment for an illness in the past 7 d among cases and matched controls stratified by zinc supplementation status. Among the 550 case-control pairs, zinc supplementation status was concordant for 418 pairs (76%). All pairs (placebo, n = 197 pairs; zinc, n = 221 pairs) were included in the analysis (Table 1). In the placebo group, cases and controls were comparable with respect to baseline characteristics and Spn carriage. However, odds of ALRI were 20 times greater among children who had received treatment for an illness 1–7 d prior to NP specimen collection [matched odds ratio (MOR) = 22.2, (95% CI, 9.06–54.4), P < 0.001]. Anecdotal evidence suggests that antibiotics are commonly used in Sarlahi to treat childhood illnesses, including ALRI and diarrhea. Treatment with antibiotics has been shown to reduce NP colonization of antibiotic-sensitive bacteria but can select for colonization with drug-resistant strains (22,23). In the zinc group, children who had 1 or more siblings <5 y of age were at higher risk for ALRI [MOR = 2.00; (95% CI, 1.20–3.34); P = 0.008]. Additionally, children who had been treated for an illness 1–7 d prior to NP specimen collection had a 16-fold greater risk of ALRI than those who had not [MOR = 16.71; (95% CI, 7.80–35.830); P < 0.001]. Conversely, Spn carriage was associated with a lower risk of ALRI [MOR = 0.61; (95% CI, 0.38–0.97); P = 0.037].
TABLE 1
Selected characteristics and Spn carriage status of ALRI cases and matched controls stratified by treatment group, Sarlahi, Nepal1
| Characteristic | ALRI cases | Controls | MOR | 95% CI | ||
|---|---|---|---|---|---|---|
| Placebo | n/total n | % | n/total n | % | ||
| Male sex | 88/197 | 44.7 | 91/197 | 46.2 | 0.93 | 0.62–1.41 |
| Child's age <12 mo | 48/197 | 25.4 | 50/197 | 24.4 | 0.67 | 0.18–2.36 |
| 1 or more siblings <age 5 y | 149/190 | 78.4 | 141/189 | 74.6 | 1.23 | 0.73–2.06 |
| Mother smokes cigarettes/bedis | 21/171 | 12.3 | 17/172 | 9.9 | 1.60 | 0.73–3.53 |
| Low socioeconomic status | 90/197 | 45.7 | 91/197 | 46.2 | 0.69 | 0.40–1.18 |
| Spn carrier | 158/197 | 80.2 | 155/197 | 78.7 | 1.11 | 0.66–1.86 |
| Treated for illness 1–7 d before NP sampling | 136/197 | 69.0 | 30/197 | 15.2 | 22.20* | 9.06–54.39 |
| Zinc | ||||||
| Male sex | 119/221 | 53.9 | 114/221 | 51.6 | 1.09 | 0.75–1.60 |
| Child's age <12 mo | 65/221 | 29.4 | 61/221 | 27.6 | 3.00 | 0.61–14.87 |
| 1 or more siblings <age 5 y | 172/207 | 83.1 | 154/210 | 73.3 | 2.00* | 1.20–3.34 |
| Mother smokes cigarettes/bedis | 32/197 | 16.2 | 28/206 | 13.6 | 1.20 | 0.67–2.17 |
| Low socioeconomic status | 84/221 | 38.0 | 83/221 | 37.6 | 0.94 | 0.57–1.54 |
| Spn carrier | 167/221 | 75.7 | 185/221 | 83.7 | 0.61* | 0.38–0.97 |
| Treated for illness 1–7 d before NP sampling | 138/221 | 62.4 | 28/221 | 12.7 | 16.71* | 7.80–35.83 |
Stratified multivariate analysis.
Adjusting for recent treatment and sibling status, zinc supplementation modified the association between Spn carriage and ALRI (Table 2; test for interaction, P = 0.097). Whereas Spn carriage was associated with a 3-fold increase in ALRI risk [adjusted MOR (AMOR) = 2.57; (95% CI, 1.42–5.96); P = 0.025] among pairs in the placebo group, Spn carriage and ALRI in the zinc group were not associated. The treatment effect did not differ between serotypes included in the current 7-, 10-, and 13-valent Spn conjugate vaccines and those that are not.
TABLE 2
Association between pneumococcal carriage and risk of ALRI stratified by treatment group, Sarlahi, Nepal12
| Factor | ALRI cases | Controls | MOR (95% CI) | AMOR3 (95% CI) | ||
|---|---|---|---|---|---|---|
| All ALRI cases3 | n/total n | % | n/total n | % | ||
| Placebo Spn carrier | 158/197 | 80.2 | 155/197 | 78.7 | 1.11 (0.66–1.87) | 2.57 (1.42–5.96)* |
| Zinc Spn carrier | 167/221 | 75.6 | 185/221 | 83.7 | 0.61 (0.38–0.97) | 0.95 (0.48–1.91) |
| Symptomatic ALRI cases at time of NP specimen collection4 | ||||||
| Placebo Spn carrier | 40/43 | 93.0 | 30/43 | 69.8 | 12.00 (1.56–92.29)* | 78.09 (3.49–1748.63)* |
| Zinc Spn carrier | 44/52 | 84.6 | 45/52 | 86.4 | 0.85 (0.29–2.55) | 2.77 (0.42–18.19) |
When we repeated this analysis for the 95 cases who were symptomatic during NP specimen collection and their matched controls, we detected evidence of a strong interaction between zinc status and Spn colonization (Table 2; test for interaction, P = 0.036). In the placebo group, the odds of ALRI were 78 times greater in Spn carriers than in noncarriers [AMOR = 78.09; (95%CI, 3.40–1748.6); P = 0.006]. By comparison, there was no significant difference in the risk of ALRI by Spn carriage status in the zinc group [AMOR = 2.77; (95% CI, 0.42–18.19); P = 0.288].
We also assessed the effect of zinc supplementation on the risk of carriage-associated ALRI, a proxy for bacterial infection. Spn carriage was detected in 422 (77%) of the 550 ALRI cases included in the initial analysis. ALRI cases and controls were stratified by the carriage status of the case and compared by treatment group (Table 3). There was insufficient evidence of an interaction between treatment status and the Spn carriage status of ALRI cases (P = 0.277). Zinc prophylaxis was associated with a nonsignificant, decreased risk of ALRI with Spn carriage [AMOR = 0.69; (95% CI, 0.41–1.17); P = 0.169], but there was no evidence that zinc decreased the risk of ALRI in the absence of Spn carriage. These results are consistent with the findings from the stratified analyses presented earlier.
TABLE 3
Association between zinc supplementation and risk of ALRI stratified by Spn carriage status of cases1
| Risk factors | ALRI cases | Controls | MOR (95% CI) | AMOR2 (95% CI) | ||
|---|---|---|---|---|---|---|
| ALRI cases: Spn carriage absent | n/total n | % | n/total n | % | ||
| Zinc | 72/128 | 56.3 | 71/128 | 55.5 | 1.06 (0.55–2.05) | 1.24 (0.41–3.45) |
| ALRI cases: Spn carriage present | ||||||
| Zinc | 209/422 | 49.5 | 222/422 | 52.6 | 0.19 (0.51–1.14) | 0.69 (0.41–1.17) |
Discussion
In this prospective, population-based, matched case-control study, daily zinc supplementation modified the association between NP Spn carriage and risk of ALRI in young Nepalese children. Spn carriage significantly increased the risk of ALRI in the placebo group, but carriage was not associated with increased risk of ALRI in the zinc group. Our finding of an interaction between zinc and Spn carriage is supported by the results of our subanalysis of the effect of zinc supplementation on risk of ALRI with and without Spn carriage. Zinc prophylaxis was inversely associated with risk of carriage-positive ALRI, whereas the association with carriage-negative ALRI was not.
Our findings suggest that zinc prophylaxis may improve immunity to Spn infections, including pneumonia, among children colonized with Spn. Bacterial carriage is a necessary, but not a sufficient, factor for the initiation of bacterial infection. The data suggest that zinc may interfere with the process of infection. Zinc deficiency impairs antibody-mediated responses to infection, which play important roles in inhibiting the colonization and phagocytosis of encapsulated bacteria (18,24). Experimental studies suggest that suboptimal zinc status damages the nonspecific barrier function of airway epithelial cells, which are responsible for clearing aspirated bacteria from the respiratory tract (25,26). In addition, results from a small supplementation trial showed that zinc prophylaxis reduces the incidence of Spn tonsillitis in adult patients with sickle cell anemia, suggesting a restoration of local antibody-mediated responses and the barrier function of the respiratory mucosa (27). The prevalence of bacterial carriage was comparable between cases and controls, and between treatment groups; thus, our data do not support the notion that zinc may decrease bacterial disease transmission by lowering the risk of bacterial carriage.
Because our results indicate that zinc is only protective against bacterial carriage-related ALRI, they raise the possibility that the protective effect of zinc on severe pneumonia may differ by etiology. Although many bacterial and viral pathogens cause similar signs and symptoms of ALRI, they differ in their pathogenesis. Thus, the differences in geographic distribution of ALRI pathogens may partially explain the variation in treatment effects in the other randomized trials evaluating the impact of zinc supplementation on the incidence of ALRI or pneumonia. We are not aware of any published data on bacterial NP carriage or on ALRI etiology from recent zinc trials; however, our findings appear to be compatible with studies by Brooks et al. (12) and Bhandari et al. (13). The results of these trials indicated that zinc had a protective effect against pneumonia, which was defined using criteria more consistent with severe disease more likely to be of bacterial etiology. It is plausible that the differences in effect size between the 2 trials could be due to differences in Spn prevalence. In the other trials, the definition was broader, making it difficult to discern if the effect of zinc differed by severity of infection. One notable difference between the trials that may also account for variations in study outcomes is that the zinc dosage for infants was 70 mg/wk in the trials that showed an overall protective effect compared with 35 mg/wk in the studies that showed no impact. Additionally, baseline zinc status was not recorded in all studies and it is possible that zinc may benefit only those who are zinc deficient.
In addition, treatment for illness within the previous 7 d was strongly associated with ALRI. Parents were not asked why they sought treatment for their child during the recall period or what form of treatment was received. One possible explanation for our finding is that early symptoms of ALRI may be easily recognized and perceived to be serious, thus warranting treatment. There is some published data from studies in South Asia to support this notion (28,29). This finding warrants further investigation, because it has implications for improving appropriate health care seeking behaviors in poor, at-risk communities.
A potential limitation of this study is the broad definition of ALRI that was used. We did not collect any data that would allow us to classify infections with any accuracy into severe and nonsevere categories, nor were blood cultures used to determine the etiology of infection. A wide variety of pathogens can cause ALRI, but most moderate to severe pneumonias tend to be the result of bacterial infections (2,30). This may partially explain why we saw no overall treatment effect of zinc but strong evidence of interaction with Spn carriage. Another issue is related to the timing of NP specimen collection among cases. NP specimens were frequently collected when cases no longer experienced signs and symptoms of infection. Our data show that the association between Spn carriage and risk of ALRI is dependent on whether the case was symptomatic at the time of swab collection. This is consistent with evidence showing that colonization rates tend to be higher during respiratory tract infections and otitis media (4,31). Thus, it is possible that we underestimated the overall effect of zinc on the relationship between Spn carriage and ALRI.
In summary, we found evidence that zinc supplementation may significantly weaken the association between Spn carriage and the risk of ALRI in young children living in areas of endemic zinc deficiency. The effect was restricted to the risk of Spn carriage-related ALRI. These findings may suggest that zinc helps to impede the infection process rather than to block NP carriage. In addition, its role in ALRI prevention may depend on etiology of infection. Differences in the prevalence of these types of infections may partially explain variations in treatment effect between recent trials, along with dissimilarities in zinc deficiency status, dosing, and ALRI definitions. Given the mounting problem of antimicrobial resistance, the slow introduction of Spn vaccines in developing countries, and the low cost of zinc, it is important that we clarify the role of zinc in the prevention of severe pneumonias both in the absence of and as an adjunct to vaccination. Further longitudinal studies are needed to evaluate the independent effect of zinc on NP carriage of bacterial pathogens, its contribution to herd immunity as well as its impact on the incidence of severe pneumonias, and to determine whether its impact differs by bacterial etiology.
Acknowledgments
We thank Dr. Reba Kanungo, Professor and Head, Department of Clinical Microbiology, Pondicherry Institute of Medical Sciences, Pondicherry, India for her assistance with the training of the laboratory staff.
Notes
1Supported by grants from the NIH, Bethesda, MD (HD 38753), the Bill & Melinda Gates Foundation, Seattle, WA (810-2054), and a cooperative agreement between Johns Hopkins University and the Office of Health and Nutrition, US Agency for International Development, Washington, DC (HRN-A-00-97-00015-00). Dr. Coles received funding support from a NIH Mentored Research Scientist Development award (K01DK07578).
2Author disclosures: C. Coles, J. Sherchand, S. Khatry, J. Katz, S. LeClerq, L. Mullany, and J. Tielsch, no conflicts of interest.
6Abbreviations used: ALRI, acute lower respiratory tract infection; AMOR, adjusted matched odds ratio; MOR, matched odds ratio; NNIPS4, Nepal Nutrition Intervention Project, Sarlahi-4; NP, nasopharyngeal; Spn, Streptococcus pneumoniae; VDC, village development committee.