Pneumococcal Carriage at Age 2 Months Is Associated with Growth Deficits at Age 6 Months among Infants in South India

Abstract

Nasopharyngeal colonization is the first step in the pathway to Streptococcus pneumoniae (Spn) infection, a leading cause of childhood morbidity and mortality. We investigated the effect of Spn colonization at ages 2 and 4 mo on growth at age 6 mo among 389 infants living in rural South India by using data from an Spn carriage study nested within a randomized, double-blind, placebo-controlled community trial designed to evaluate the impact of newborn vitamin A supplementation on Spn carriage in the first 6 mo of life. Primary outcomes were weight, length, and anthropometric indices of nutritional status. Growth data at age 6 mo were available for 84% (389 of 464) of infants in the Spn carriage study. Carriage at age 2 mo was associated with increased odds of stunting [OR: 3.07 (95% CI: 1.29, 7.36) P = 0.012] and lower weight [β: −266 g (95% CI: −527, −5) P = 0.045], length [β: −1.31 cm (95% CI: −2.32, −0.31) P = 0.010], and length-for-age Z scores [β: −0.59; (95% CI: −1.05, −0.13) P = 0.012] at age 6 mo. Spn carriage at age 4 mo did not affect growth. Carriage of invasive serotypes at age 2 mo was associated with decreases in mean weight [β: −289 g; (95% CI: −491, −106) P = 0.002] and length [β:−0.38 cm (95% CI: −1.49, −0.01) P = 0.047] at age 6 mo. Newborn vitamin A supplementation did not modify the association between Spn carriage and growth. Results suggest that pneumococcal carriage at age 2 mo is an independent risk factor for poor growth in young infants. Future studies need to clarify the role of Spn carriage on growth retardation in low-income countries.

Introduction

Globally, Streptococcus pneumoniae (Spn)⁸ infections remain a leading cause of pediatric morbidity and mortality (1, 2). It is estimated that Spn is responsible for ~800,000 deaths and nearly 15 million severe infections, including pneumonia, annually in children aged <5 y, with >90% of cases occurring in low-income countries (1). Asymptomatic nasopharyngeal colonization with Spn is considered to be the first step in the pathway to infection (3, 4). Although carriage, which reflects the ongoing colonization of the host, is relatively common, only a small proportion of colonized children go on to develop pneumococcal disease (2, 3). Risk of Spn infection is greatest soon after a serotype is acquired or first establishes itself within the host’s nasopharynx (5–8). Poor growth and undernutrition are known to exacerbate the risk and severity of infections (2, 9, 10). In contrast to industrialized countries, children in low-income settings experience early and abundant Spn carriage well before the age of 6 mo, which may partially explain why the incidence of Spn infections is greatest in these countries (11–14). In turn, infection can adversely affect nutritional status and increase risk of future infections (9). Pneumococcal conjugate vaccines help to break this cycle by preventing invasive infections and by reducing acquisition of vaccine serotypes, which decreases disease transmission (15–17).

Research on Spn carriage is focused primarily on its role as a risk factor for Spn infection. However, emerging evidence suggests that Spn carriage may have direct and detrimental effects on child health, unrelated to its role in Spn infection. Results from a study by Sleeman et al. (18) suggested that long-term Spn carriage is associated with increased risk of mild upper respiratory illness, possibly of viral etiology. In addition, Spn carriage may adversely affect growth and nutritional status in early childhood. Findings from an Spn carriage study in Vellore, India (19), indicate that previous carriage may have a detrimental effect on the growth of young children from a population at risk for undernutrition. The study assessed risk factors for Spn colonization in addition to effects of acquisition and colonization on morbidity and growth in infants followed from birth to age 12 mo. Study results showed a significant association between Spn carriage and poor growth in the study population. The duration of carriage was associated with significant decreases in length and weight gain.

There is little published information on the impact of carriage of respiratory pathogens on childhood growth; however, findings from the Vellore study are consistent with the results of studies of carriage of enteric pathogens, including Helicobacter pylori (20) and Cryptosporidium parvum (21), which are associated with long-lasting deficits in growth in the absence of symptomatic illness. If Spn carriage contributes to growth retardation and worsens nutritional status in early childhood, it may indirectly increase the risk of other life-threatening infections. Whether the potential effect on growth is dependent on the age of carriage or is influenced by Spn serotype has not been studied. The effects of vaccines on Spn acquisition may mitigate the adverse effects on growth depending on the local prevalence of serotypes included in the licensed 13-valent Spn conjugate vaccine.

We investigated the effect of Spn colonization at ages 2 and 4 mo on growth and nutritional indices at age 6 mo in a cohort of 389 infants who participated in a newborn vitamin A (VA) supplementation trial in South India (22). We also assessed whether the association between Spn carriage and growth differed between vaccine and nonvaccine Spn serotypes.

Participants and Methods

Study population.

The participants of the InPACT (Infant Pneumococcal Acquisition/Colonization in Tamil Nadu) study were recruited from an ongoing VA supplementation study in newborns [Vitamin A Supplementation In Newborns (VASIN) trial]. The VASIN was a 3-y, randomized, double-blind, placebo-controlled trial that aimed to evaluate the impact of VA given to the infants at birth on mortality, morbidity, and growth through the age of 6 mo. This trial was conducted in 2 rural areas in southern India, Natham and Karriyapatty, in the state of Tamil Nadu. Both target areas were representative of rural South Asian communities that experienced endemic VA deficiency and a high incidence of acute respiratory infections. The findings of the VASIN trial have been published previously (22, 23).

InPACT enrollment.

The InPACT study design and methodology and findings were previously described (24). Briefly, infants who were born into the VASIN trial, who survived to the age of 2 to 2.5 mo, and who lived in 1 of the 8 supervisory areas selected from the Natham block were eligible to be in the InPACT study. The Natham block was targeted because of its larger and denser population compared with Karriyapatty, and the 8 supervisory areas had the highest birth rates. VASIN trial records were used to recruit eligible infants, and parents gave consent for voluntary participation. A total of 464 infants were included in the study. Nasopharyngeal samples were obtained during follow-up from October 1998 through June 1999. The results of the InPact trial showed that newborn VA supplementation delayed the age of initial Spn acquisition in young infants when supplemented at birth (24, 5). The risk of Spn carriage in infants aged 4 mo who were not colonized by age 2 mo was significantly reduced in the VA group compared with the placebo group [OR: 0.51 (95% CI: 0.28, 0.92) P = 0.02]. In addition, the odds of colonization tended to be lower in the treatment group than in the placebo group at age 4 mo [OR: 0.73 (95% CI: 0.48, 1.1) P = 0.13].

Ethical review.

The study protocols for both VASIN and InPACT were approved by the Ethical Committee of the Aravind Eye Hospital and Institute, the Tamil Nadu government, and the Committee on Human Research of the Johns Hopkins University Bloomberg School of Public Health. Oral informed consent from parents or guardians was obtained because of the low literacy levels in the study location.

Data collection.

Demographic information and household-level data were obtained from the VASIN trial. This included data collected by field staff at the time of delivery, such as infant weight, gender, and colostrum feeding and the mother’s history of night blindness during pregnancy. Five trained field workers also collected nasopharyngeal samples from all InPACT study infants at the age of 2, 4, and 6 mo. Weight measurements were obtained by using an electronic infant weighing scale (Seca model 727; Seca). All anthropometric measurements were performed by trained field staff members with the use of a standardized protocol. Two staff members were assigned to measure weight and recumbent length. Newborn weight as well as weight at age 6 mo were measured by using a Seca model 727 electronic infant weighing scale. Upon arrival at the site where the child was weighed, the scale was removed from its case and placed on a flat and level surface. The scale was turned on and when the screen read “0.0,” the undressed child was placed gently on the scale. Weight was measured to the nearest 2 g once the screen was steady for at least 3 s. Scales were calibrated daily against a known standard weight. Recumbent length was measured 3 times to the nearest 0.1 cm on a wooden length board (Shorr Infant Measuring Board; Shorr Productions). Before each measurement, the team ensured that the infant’s knees were locked and that his or her heels were firmly against the footboard. For length, the median of 3 values was used in analyses.

Nasopharyngeal specimen collection required inserting a small, flexible rayon-tipped swab (DIFCO CultureSwab Transport System with Amies Media; BD Diagnostics) into the posterior nasopharynx for ≥5 s or rotating the swab 180° before removal. All newly collected samples were placed into Amies transport medium and taken to the microbiology laboratory housed in the Aravind Eye Hospital within 10 h. Samples not collected within a window of 2 to 14 d of the designated collection date were considered a “missing visit.”

Laboratory procedures.

All nasopharyngeal swabs were inoculated onto tryptic soy agar plates (Becton Dickinson) containing 5% sheep blood and 5 mg/L gentamicin (Nathan Pirumal) within 12 h of arrival at the microbiology laboratory. After inoculation, plates were incubated at 37°C in 5% CO₂ for 18 to 24 h. Optochin (Taxo) inhibition and bile solubility (Himedia) tests were used to confirm growth of pneumococcal colonies. Optochin inhibition zones >13 mm were considered “culture positive,” whereas those with 9–13-mm zones were “culture negative.” For “culture indeterminate” samples, the bile solubility test was used to confirm the presence of pneumococci. Infants with culture-positive or bile solubility test-confirmed samples were identified as pneumococci carriers. A pneumococcal reference strain (5603; American Type Culture Collection) was used to ensure quality control. Pneumococcal isolates were serotyped and grouped by using PNEUMOTEST kits (Staten Serum Institute), which included antisera of serotypes or groups contained in the licensed 23-valent pneumococcal polysaccharide vaccine (25). The antisera included in the kits react 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. The kits use different combinations of type-specific pneumococcal rabbit antisera to identify 23 pneumococcal serogroups or serotypes of strains most frequently associated with invasive disease. Serotyped isolates were further classified based on whether or not they were included in the Prevnar 13 vaccine (13vPnC; Wyeth Pharmaceuticals, a subsidiary of Pfizer, Inc.), which is the pneumococcal conjugate vaccine that was licensed in the United States in 2010 and which is undergoing evaluation for use in low-income countries. The vaccine includes serotypes 1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F, and 23F. These serotypes are reported to account for 80 to 92% of the invasive pneumococcal infections that occur each year in children <5 y of age (26).

Definition and measurement of outcomes.

The dichotomous outcomes assessed in this study were the prevalence of underweight, stunting, and wasting at 6 mo of age. Underweight, stunting, and wasting were defined as below −2 SD from the median weight-for-age, length-for-age, and weight-for-length of the reference population according to WHO Child Growth Standards (27). The continuous outcomes were weight, length, weight-for-age Z score (WAZ), length-for-age Z score (LAZ), and weight-for-length Z score (WLZ) at age 6 mo.

Statistical analyses.

Two-tailed χ² tests and t tests were used to explore the distribution of baseline characteristics (Stata 11.0; Stata Corporation). Similar methods were used to compare infants with anthropometric data to those infants for whom these data were missing. Bivariate analyses were conducted by using 2-tailed χ² tests to determine the association between growth outcomes at age 6 mo and covariates of interest. In the multivariate linear and logistic regression models, the effect of Spn carriage at ages 2 and 4 mo on growth at age 6 mo was adjusted for covariates that were statistically associated (P ≤ 0.1) in the bivariate analyses. OR 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. Because of the characteristics of pneumococcal transmission and the periodic nature of specimen collection during this InPACT study, the estimates obtained for pneumococcal colonization likely underestimate the true incidence in this population. Values in the text are means ± SD unless indicated otherwise.

Results

Enrollment and follow-up.

Four hundred sixty-four infants were enrolled in the InPACT study at the age of 2 mo. Of these, 4 died and 7 were lost to follow-up (Supplemental Fig. 1). Weight and length data were not collected for 64 of the infants for the following reasons: family migration from study area (45.3%), infant was beyond the eligible age window in which to collect anthropometric data (31.3%), infant was adopted (1.6%), missing anthropometric data form (20.3%), and family refusal to participate (1.6%) (Supplemental Fig. 1). Anthropometric data at age 6 mo were available for 389 and 355 of the InPACT infants with colonization data at ages 2 and 4 mo, respectively. In total, there were 354 infants with colonization and anthropometric data at both time points. Infants who were eligible to be included in this analysis were generally comparable to those who did not have anthropometric data at age 6 mo. The proportion of infants in each group was similar with regard to gender, low birth-weight status, and level of maternal education. Mean gestational age at the time of delivery was not different across groups. However, mean birth weight for infants for whom anthropometric data at age 6 mo were not available tended to be lower than that of children with growth data at age 6 mo (2584 ± 409 vs. 2682 ± 424 g, P = 0.07). In contrast, the proportion of infants receiving colostrum feeding tended to be greater in infants without growth data at age 6 mo (89.2%) compared with those with 6-mo data (80.5%; P = 0.08).

Baseline characteristics.

Approximately 56% (219 of 390) of the infants were male. In addition, nearly one-third of the infants weighed <2500 g at birth (Table 1). All infants were breast-fed, and 80% received colostrum. All infants came from families with low socioeconomic status. The vast majority of the infants (94.6%) resided in homes that used wood or other biomass as the primary cooking fuel, and 45% of mothers had no educational experience. Approximately 68% of the infants received VA supplements soon within 48 h of birth as part of the parent trial.

TABLE 1.

Characteristics of InPACT study infants included in analysis at ages 2 mo (n = 389)¹

Baseline characteristics	Spn carriage at age 2 mo
	n (%)
Vitamin A treatment
Yes	201 (51.7)
No	188 (48.3)
Gender
Male	219 (56.3)
Female	170 (43.7)
Low birth weight (<2500 g)
Yes	117 (30.2)
No	271 (69.9)
Colostrum fed to newborn
Yes	313 (80.1)
No	76 (19.5)
Household cooking fuel
Wood	368 (94.6)
Other	21 (5.4)
Maternal education
None	216 (55.5)
≥1 y	173 (44.5)
Time birth weight assessed
<48 h	265 (68.1)
48–72 h	31 (8.8)
>72 h	87 (21.1)

¹InPACT, Infant Pneumococcal Acquisition/Colonization in Tamil Nadu; , Streptococcus pneumoniae.

Prevalence of selected growth outcomes.

Weight and length were measured at 6 mo of age. The prevalences of stunting, wasting, and underweight at age 6 mo were 25.6, 21.3, and 37.4%, respectively. Body weight was 6364 ± 843 g, and length was 64.0 ± 3.1 cm. WAZ, LAZ, and weight-for-length Z score were −1.67 ± 1.09, −1.28 ± 1.39, and −1.06 ± 1.33, respectively.

Effect of nasopharyngeal carriage of pneumococci on growth.

Initially, we conducted bivariate analyses to estimate the association between dichotomous growth outcomes and 4 categories of Spn carriage: never colonized, carriage at age 2 mo only, carriage at age 4 mo only, and carriage at ages 2 and 4 mo. The odds of stunting at age 6 mo were ~2-fold greater among infants who were only colonized at age 2 mo than among noncarriers, and this was the only association to approach significance (Table 2). Neither carriage at age 4 mo nor carriage at ages 2 and 4 mo was associated with stunting. Risk of underweight and wasting at age 6 mo was not associated with any carriage category. In the multivariate logistic and linear regression models, the effect of Spn carriage on growth outcomes was adjusted for gender, birth weight, colostrum feeding, maternal education, and gestational age. Among the categorical growth indices, only the association between stunting and Spn carriage at age 2 mo alone was significant. The odds of stunting were 3 times greater among infants who experienced carriage at 2 mo of age compared with infants who were not colonized at 2 mo [OR: 3.07 (95% CI: 1.29, 7.36) P = 0.012]. On the contrary, carriage at 4 mo and at ages 2 and 4 mo was not associated with underweight, stunting, or wasting. We then assessed the relationship between continuous growth outcomes and carriage status by using multiple linear regression analysis. Carriage at 2 mo alone was associated with a 266-g decrease in mean weight and a 1.31-cm decrease in mean length at age 6 mo. In addition, there was a 0.35 decrease in WAZ, and a 0.59 decrease in LAZ (Table 3). Carriage at age 4 mo alone, and carriage at ages 2 and 4 mo had no impact on growth indices after adjustment for the effects of other confounders and covariates. It should be noted that VA supplementation did not affect growth at age 6 mo in the parent trial (22). In addition, carriage prevalence and duration of diarrhea and acute respiratory illness episodes were evaluated as potential confounders and effect modifiers of the association between Spn carriage and growth; however, no evidence of confounding or interaction was detected.

TABLE 2.

Association between Spn carriage on dichotomous growth indices at age 6 mo in infants in South India (InPACT study)¹

	Underweight (WAZ < −2 SD)			Stunting (LAZ < −2 SD)			Wasting (WLZ < −2 SD)
Spn carriage pattern	n (%)	Crude OR (95% CI)	Adjusted OR2 (95% CI)	n (%)	Crude OR (95% CI)	Adjusted OR2 (95% CI)	n (%)	Crude OR (95% CI)	Adjusted OR2 (95% CI)
Never colonized	68 (33.8)	1	1	68 (26.5)	1	1	68 (26.5)	1	1
Carriage at age 2 mo only	58 (43.1)	1.48 (0.72, 3.05)	1.81 (0.80, 4.10)	58 (41.4)	1.96 (0.93, 4.15)	3.07 (1.29, 7.36)	58 (20.7)	0.72 (0.32, 1.67)	0.83 (0.34, 1.99)
Carriage at age 4 mo only	101 (38.6)	1.23 (0.65, 2.34)	1.48 (0.73, 3.01)	101 (18.8)	0.64 (0.31, 1.34)	0.76 (0.34, 1.70)	101 (20.8)	0.73 (0.35, 1.50)	0.82 (0.39, 1.71)
Carriage at ages 2 and 4 mo	127 (35.4)	1.07 (0.58, 2.00)	1.44 (0.70, 2.93)	127 (22.8)	0.82 (0.42, 1.62)	1.24 (0.57, 2.71)	127 (21.3)	0.75 (0.38, 1.49)	0.90 (0.43, 1.87)

¹LAZ, length-for-age Z score; InPACT, Infant Pneumococcal Acquisition/Colonization in Tamil Nadu; WAZ, weight-for-age Z score; WLZ, weight-for-length Z score; , Streptococcus pneumoniae.

²Adjusted for the effects of gender, birth weight, colostrum feeding, gestational age, and maternal education.

TABLE 3.

Association between Spn carriage and continuous measures of growth at age 6 mo in infants in South India (InPACT study)¹

Spn carriage pattern	Weight (g)		Length (cm)		WAZ		LAZ		WLZ
	β (95% CI)	P	β (95% CI)	P	β (95% CI)	P	β (95% CI)	P	β (95% CI)	P
Never colonized	0		0		0		0		0
Colonized at age 2 mo only	−266 (−527, −5)	0.045	−1.31 (−2.32, −0.31)	0.010	−0.35 (−0.70, −0.00)	0.050	−0.59 (−1.05, −0.13)	0.012	−0.06 (−0.43, 0.54)	0.824
Colonized at age 4 mo only	−71 (−296,153)	0.535	−0.32 (−1.18, 0.54)	0.468	−0.08 (−0.38, 0.22)	0.594	−0.13 (−0.52, 0.27)	0.536	−0.00 (−0.42, 0.42)	0.995
Colonized at ages 2 and 4 mo	−163 (−385, 58)	0.148	−0.51 (−1.36, 0.38)	0.1244	−0.20 (−0.50, 0.10)	0.191	−0.21 (−0.61, 0.18)	0.294	−0.04 (−0.45, 0.38)	0.869

¹Values were adjusted for the effects of gender, birth weight, colostrum feeding, gestational age and maternal education. LAZ, length-for-age Z score; InPACT, Infant Pneumococcal Acquisition/Colonization in Tamil Nadu; WAZ, weight-for-age Z score; WLZ, weight-for-length Z score; , Streptococcus pneumoniae.

As noted earlier, results from the carriage trial indicated that newborn VA supplementation delayed the age of Spn acquisition (24). Given these findings, we assessed whether VA supplementation status modified the observed association between growth deficits at age 6 mo and early Spn carriage. We detected no evidence of effect modification of VA supplementation on the association between outcomes of interests at age 6 mo and Spn carriage at ages 2 and 4 mo.

Impact of colonization with invasive Spn serotypes at age 2 mo on growth at age 6 mo.

On the basis of our observations that Spn carriage at age 2 mo was associated with growth deficits at age 6 mo, we assessed whether the effect on growth of Spn colonization at age 2 mo differed by serotype invasiveness. Invasive carriage was defined as carriage of ≥1 of the 13 serotypes included in the 13vPnC pneumococcal conjugate vaccine. We assumed 100% cross-protection between serotypes within serogroups. Bivariate analyses showed that infants who experienced invasive carriage at age 2 mo had almost 2 times greater odds of being stunted than did infants who were not infected with invasive serotypes [OR: 1.93 (95% CI: 1.08, 3.46) P = 0.027] (Table 4). There were no significant associations between invasive carriage and underweight or wasting. In the multivariate logistic regression model, carriage of invasive serotypes remained associated with stunting at age 6 mo [OR: 2.16 (95% CI: 1.11, 4.20) P = 0.024]. In the multiple linear regression analyses, mean weight at 6 mo was 298 g less for infants who were colonized with invasive serotypes compared with those who were not colonized (Table 5). We also observed a mean 0.75 cm decrease in length at age 6 mo compared with that of noncarriers. Similarly, WAZ and LAZ were lower among infants carrying invasive serotypes compared with those who were not colonized. There was no apparent association between invasive carriage and wasting. Carriage of invasive serotypes at age 4 mo or at ages 2 and 4 mo was not associated with any growth outcome.

TABLE 4.

Association between Spn serotype carriage pattern at age 2 mo on dichotomous growth indices at age 6 mo in infants in South India (InPACT study)¹

	Underweight (WAZ < −2 SD)			Stunting (LAZ < −2 SD)			Wasting (WLZ < −2 SD)
Spn serotype carriage pattern	n (%)	Crude OR (95% CI)	Adjusted OR2 (95% CI)	n (%)	Crude OR (95% CI)	Adusted OR2 (95% CI)	n (%)	Crude OR (95% CI)	Adjusted OR2 (95% CI)
No carriage	186 (21.5)	1	1	186 (36.6)	1	1	186 (22.0)	1	1
Carriage of non-13vPnC strain	125 (26.5)	0.9 (0.57, 1.46)	1.11 (0.65, 1.88)	125 (34.4)	1.31 (0.77, 2.22)	1.73 (0.95, 3.13)	125 (17.6)	0.76 (0.42, 1.34)	0.87 (0.48, 1.59)
Carriage of 13vPnC strain	78 (34.6)	1.41 (0.83, 2.42)	1.51 (0.82, 2.76)	78 (44.9)	1.93 (1.08, 3.46)	2.16 (1.11, 4.20)	78 (25.6)	1.22 (0.66, 2.26)	1.35 (0.70, 2.58)

¹13vPnC serotypes: 1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F, and 23F. Analysis was based on pneumococcal serogroup designation. LAZ, length-for-age Z score; InPACT, Infant Pneumococcal Acquisition/Colonization in Tamil Nadu; WAZ, weight-for-age Z score; WLZ, weight-for-length Z score; , Streptococcus pneumoniae.

²Adjusted for the effects of gender, birth weight, colostrum feeding, gestational age, and maternal education.

TABLE 5.

Association between Spn serotype carriage at age 2 mo and continuous measures of growth at age 6 mo in infants in South India (InPACT study)¹

Spn serotype carriage pattern	Weight (g)		Length (cm)		WAZ		LAZ		WLZ
	β (95% CI)	P	β (95% CI)	P	β (95% CI)	P	β (95% CI)	P	β (95% CI)	P
No carriage	—	—	—	—	—	—	—	—	—	—
Carriage of non-13vPnC strain	−96 (−262, 70)	0.257	−0.53 (−1.17, 0.11)	0.265	−0.12 (−0.35, 0.10)	0.295	0.24 (−0.54, 0.05)	0.311	0.09 (0.22, 0.40)	0.578
Carriage of 13vPnC strain	−298 (−491, −106)	0.002	−0.75 (−1.49, −0.01)	0.047	−0.38 (−0.54, −0.13)	0.004	−0.33 (−0.67, 0.00)	0.054	−0.21 (−0.57, 0.14)	0.250

¹13vPnC serotypes: 1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F, and 23F. Analysis assumes complete cross-protection between serotypes within serogroups. Values were adjusted for the effects of gender, birth weight, colostrum feeding, gestational age, and maternal education. LAZ, length-for-age Z score; InPACT, Infant Pneumococcal Acquisition/Colonization in Tamil Nadu; WAZ, weight-for-age Z score; WLZ, weight-for-length Z score; , Streptococcus pneumoniae.

Discussion

In this prospective, population-based study we observed that nasopharyngeal colonization with Spn at age 2 mo was associated with significant reductions in growth and nutritional status at age 6 mo, after the effects of gender, birth weight, colostrum feeding, maternal education, and gestational age at birth were controlled for. Colonized infants were at increased risk of stunting and had decreased LAZ and WAZ at age 6 mo compared with noncolonized children. Among continuous outcomes, weight and length at age 6 mo were significantly lower in Spn carriers than in noncarriers. In contrast, Spn carriage at age 4 mo and at ages 2 and 4 mo was not associated with growth outcomes at age 6 mo.

Our results are compatible with evidence from other small-scale studies that assessed the relationship between growth and pathogen colonization in young children, despite some differences in outcomes definition and timing. Results from an Spn colonization study conducted in Vellore, India (19), showed that Spn carriage reduced growth in infants who were followed from birth through age 12 mo and in whom carriage status was assessed every 3 wk. The mean difference in WAZ and LAZ between colonized and uncolonized infants was −0.05 and −0.08, respectively, for each month of Spn colonization. Thus, 4 mo of Spn carriage would be estimated to result in a decrease of 0.20 and 0.32 in WAZ and LAZ, respectively, between colonized and uncolonized children. In addition, the investigators reported differences in weight and length gain of −48 g and −0.22 cm per month of colonization, respectively. Weight and length gain in children colonized for 4 mo would be 200 g less and would be nearly 1 cm shorter than that of uncolonized children. Our estimates of the magnitude of the effect of Spn colonization on growth were consistently greater than for the Vellore study. In contrast to our study, the Vellore study focused on Spn carriage duration and its impact on growth at age 12 mo. The results were averaged over the first year of life and did not assess the influence of age of colonization on outcomes or consider changes in growth velocity during this period. In addition, previous studies have shown that Spn carriage prevalence is greater among young children in rural South India compared with those in Vellore (8, 24). These differences may partially account for the differences in effect size between studies.

Our findings are also comparable to results from studies that evaluated the effect of colonization with enteropathogens on childhood growth. Thomas et al. (20) examined the association between gastric colonization with H. pylori and growth in Gambian infants. Their studies showed WAZ at age 6 mo were 0.48 lower among infants colonized at age 3 mo compared with noncolonized infants. In addition, they reported that each month of H. pylori colonization was associated with a 0.063 decline in LAZ. A similar study conducted in Colombia showed significant, albeit smaller, mean reductions in LAZ scores (−0.042/mo) associated with H. pylori colonization in children aged 12 – 60 mo. Checkley et al. (28) evaluated the effects of C. parvum infections on growth in 207 children during the first 2 y of life. They reported that infants who were colonized with C. parvum gained 162 g less weight and were 0.84 cm shorter than noncolonized infants 1 mo after infection (21). The magnitude of the weight deficit was greater in infants than in toddlers. For these children, catch-up weight gain was completed 6 mo after infection; however, the effects on length persisted.

Taken together, the evidence suggests a strong association between early pathogen colonization and faltering growth and undernutrition. Moreover, our results suggest that the effects of Spn colonization may extend beyond the risk of pneumococcal disease. If carriage adversely affects infant growth and nutritional status, it may indirectly increase childhood susceptibility to a broad range of severe infections through the infection-undernutrition cycle (9).

The mechanism by which Spn may affect growth has not been well studied. Limited evidence from animal and in vitro studies suggests the possibility that subclinical chronic inflammation of the nasopharynx may partially explain the observed relationship between pneumococcal carriage and faltering growth (29–31). Spn carriage is associated with the release of proinflammatory cytokines, including IL-8 and interferon, and other inflammatory mediators, including macrophage inflammatory protein, and an increase in oxidative stress (30–33). Although the inflammatory state may help to prevent the establishment of infection, it does not appear to eliminate carriage, which is why the duration of carriage of the same strain can endure for weeks or even months (7, 34). It is possible that the sustained inflammatory state may exact a nutritional cost through increased metabolic demands and nutrient losses in the host. Concentrations of insulin-like growth factor-1, a hormone that plays an important role childhood growth, are inversely associated with proinflammatory cytokine levels in preterm infants (35). One potential difference between colonization of respiratory pathogens and enteropathogens is that enteropathogens, such as C. parvum, may also affect growth and nutritional status by altering permeability of the gastrointestinal tract and by blocking absorption of nutrients (36). Thus, the effect size is likely to be pathogen dependent and may be influenced by age at the time of infection and nutritional status of children before infection. Potential mechanisms of action underlying the relationship between Spn carriage and poor growth will require further study to clarify the role of colonization in childhood growth.

Although >90 pneumococcal serotypes have the capacity to cause pneumococcal infection, the risk of infection is serotype dependent. The efficacy of the current pneumococcal conjugate vaccines is based on the inclusion of 13 invasive serotypes, which are those that are most frequently responsible for pneumococcal disease. Our results indicate that invasive serotypes were associated with weight and length deficits at age 6 mo. It is possible that the observed association between invasive serotypes and growth may be related to the presence of virulence factors or that they trigger reduced innate responses, which could lead to state of subclinical chronic inflammation (37, 38). Our findings suggest that current pneumococcal conjugate vaccines could play a role in mitigating the effects of Spn carriage on growth. However, the first dose of these vaccines cannot be administered before the second month of life, which may limit their impact on growth indices. These results highlight the need for new strategies to provide protection in early infancy, either through maternal immunization and passive antibody transfer or by the development of vaccines that can be administered before the second month of life (39–41).

A limitation of this study was that the carriage data were only measured at 2, 4, and 6 mo of age. Spn carriage is a dynamic process, and it is possible that had we assessed carriage status at shorter intervals beginning at age 1 mo we may have obtained different estimates of treatment effect. However, based on the carriage studies discussed above, the effect of carriage tends to be inversely related to age. Thus, colonization at age 1 mo may have a stronger effect on growth at age 6 mo than does colonization at age 2 mo. In addition, guidelines for the detection of Spn colonization call for the use nasopharyngeal swabs. Unfortunately, this method can lack sufficient sensitivity to detect low levels of Spn carriage. Any misclassification of carriage status is likely to be nondifferential; therefore, we are likely to have underestimated the true association between Spn carriage and growth. There is a potential for selection bias because we did not did adjust our outcomes for nutritional status at age 2 mo when Spn colonization status was first assessed. However, analysis of demographic and household characteristics of the participants indicates no important differences by colonization status.

Future studies need to determine if Spn carriage has long-lasting adverse effects on weight gain and linear growth. In addition, pneumococcal vaccines, especially those that can be administered in the neonatal or immediate postneonatal period, should be evaluated for their impact on growth outcomes. Spn carriage poses a potential threat to childhood growth and development, underscoring the bidirectional relationship between infection and undernutrition and the need for integrated solutions to address both health issues. Consequently, programs that use a combination of nutritional and non-nutritional interventions targeting risk factors for poor growth (e.g., micronutrient supplementation and vaccination) are likely to be more effective in preventing child growth retardation than those that use a single approach.