Background

Pneumonia is the leading infectious cause of childhood mortality worldwide [1]. As part of the Innovative Treatments in Pneumonia (ITIP) project in Lilongwe, Malawi, a randomized controlled clinical trial (ITIP1) was conducted to evaluate treatment with amoxicillin for fast-breathing pneumonia (FBP) in children (Table 1) [2]. Clinical trials evaluating treatment for pneumonia may exclude children with comorbidities who are at high risk for mortality or have other complications. Pneumonia in children with comorbidities is common, and many factors determine whether contact with an etiologic agent will result in a severe episode of pneumonia, and whether the episode will be deadly [3,4,5,6,7]. These factors can be related to the child (e.g., sex, age, underlying diseases), to the disease (e.g., type of infection), to the environment, the family and its socio-economic status, and to the health system and type of care [8]. A systematic review and meta-analysis of risk factors for mortality from acute lower respiratory infections (ALRI) in children under five years of age in low- and middle-income countries found chronic underlying diseases (odds ratio 4.76, 95% confidence interval 3.27–6.93), HIV/AIDS (4.68, 3.72–5.90), and severe malnutrition (4.27, 3.47–5.25) were associated with ALRI mortality [8]. In an effort to better understand pneumonia treatment outcomes among higher risk African children post-introduction of Haemophilus influenzae type b and Streptococcus pneumoniae conjugate vaccines, and to interpret the results of the concurrent FBP clinical trial (ITIP1), a prospective observational study (ITIP3) was conducted to assess the clinical outcomes of children aged 2 to 59 months with both pneumonia and comorbidities [2, 9]. We compare clinical outcomes of children with FBP and selected comorbidities (ITIP3) to those without these selected comorbidities (ITIP1) (Fig. 1).

Table 1 Study definitions, eligibility criteria, and clinical outcomes
Full size table
Fig. 1
figure 1

ITIP1 and ITIP3 fast-breathing pneumonia cohort consort diagram by comorbidities

Full size image

Methods

Study design

The objective of this study was to investigate whether the percentage of children with FBP cured at Day 14 was lower among those with HIV infection or exposure, malaria, severe acute malnutrition, or anemia than among those without these selected comorbidities who were treated similarly. Children aged 2 to 59 months meeting the case definition of FBP in the pediatric outpatient departments of Kamuzu Central Hospital (KCH) and Bwaila District Hospital (BDH) in Lilongwe, Malawi were screened by study staff to determine ITIP eligibility (Table 1). Those who were excluded from the ITIP1 clinical trial because of comorbidities were assessed for enrollment into ITIP3. ITIP1 enrollment began June 7, 2016 with the last visit completed June 20, 2017. A total of 1126 children were enrolled in ITIP1 with 564 randomized to receive 3 days of oral amoxicillin, the standard of care treatment group, and 562 randomized to receive placebo [2]. ITIP3 FBP enrollment began November 8, 2016 with the last study visit completed on July 14, 2018. Per Malawian guidelines and KCH protocols, each child in the ITIP3 FBP cohort received standard of care treatment for their pneumonia, typically 5 days of oral amoxicillin, and for their comorbidities.

The study was conducted in accordance with the International Conference on Harmonisation, Good Clinical Practice and the Declaration of Helsinki 2008, and was approved by the Western Institutional Review Board in the state of Washington, USA; the College of Medicine Research and Ethics Committee, Blantyre, Malawi; and the Malawi Pharmacy, Medicines and Poisons Board. The study was registered with ClinicalTrials.gov: NCT02960919.

Study procedures

On Day 1, after enrollment informed consent was obtained from the caregiver or legal guardian by study staff, ITIP eligible children received a physical examination, and information regarding their medical and vaccination history along with additional socio-demographic information was collected. Screening and enrollment were conducted in the pediatric outpatient departments of KCH or BDH, with BDH enrollees transferred to KCH for continued evaluation, observation, and admission, if indicated. Hospital observation or admission and all follow-up visits occurred solely at KCH.

Enrolled children were treated and prospectively followed by study staff with scheduled study visits on Days 2 to 4 and 14 (all in-person) in ITIP1 and on Days 2 (if hospitalized), 6 and 14 (in-person), and 30 (by phone) in ITIP3. During in-person follow-up visits, study staff reviewed the medical history since the last study visit and performed a physical examination including respiratory rate, chest indrawing, and pulse oximetry assessments. In case of a no-show at scheduled follow-up visits, a home visit was conducted by study staff. If a phone call on Day 30 was not possible due to no phone in the home, study staff conducted a home visit to obtain Day 30 outcome information in ITIP3. More detailed study procedures for children enrolled in ITIP1 and ITIP3 have been described previously [2, 9].

Study outcomes

Primary endpoints for this study analysis included treatment failure (TF) rates and clinical outcomes of children treated for FBP and clinical cure rates at Day 14 in select subgroups defined by comorbidities (HIV infection or exposure, malaria, severe acute malnutrition, or anemia) with comparisons to children in the standard of care treatment group of ITIP1. Secondary endpoints included: treatment responses (vital signs, oxygen saturation, length of hospital stay); and proportion of children who were re-hospitalized or died.

Time points for TF assessment were chosen in accordance with current World Health Organization (WHO) pneumonia management guidelines [10]. In ITIP3, TF was assessed on Day 6 and in ITIP1, TF was assessed on or prior to Day 4. Children enrolled to ITIP3 were followed for 30 days post-enrollment. ITIP1’s clinical trial population was followed for 14 days post-enrollment. Of note, development of very fast breathing (as defined in Table 1) is included here in the TF definition for ITIP1 and ITIP3, but was not included in the TF definition for ITIP1 protocol and primary manuscript [2].

Statistical analysis

Descriptive statistics are provided for clinical outcomes. Chi-squared statistics or Fisher’s exact tests were used to compare TF and clinically cured rates between ITIP3 and ITIP1 treatment groups. Tests were performed as two-sided tests with alpha = 0.05. No adjustments were made for confounders, and no formal statistical comparisons were made by comorbidity subgroup due to the small number of ITIP3 FBP children. No adjustments were made for multiple comparisons because of the observational and exploratory nature of this study.

Results

Of the 79 ITIP3 FBP children enrolled, 72.2% (57/79) were HIV-infected or exposed, 25.3% (20/79) had malaria, 2.5% (2/79) met severe acute malnutrition criteria either by mid-upper arm circumference < 115 mm or weight-for-height z-score < − 3, and 21.5% (17/79) had anemia defined as a hemoglobin < 8.0 g/dL (Fig. 1). ITIP3 TF assessment data and clinical outcome by Day 14 were available for 71 (89.9%) children. ITIP1 TF assessment data by Day 4 was available for 552 and 543 children in the amoxicillin and placebo groups, respectively, and ITIP1 clinical outcome data by Day 14 was available for 543 and 527 children in the amoxicillin and placebo groups, respectively.

When comparing ITIP3’s FBP cohort to all children enrolled in ITIP1’s FBP clinical trial, there were a few differences in baseline characteristics (Table 2). Children with FBP in ITIP3 were significantly younger than those in ITIP1, with 54.4% (43/79) children 2–11 months of age compared to 34.8% (392/1126) in ITIP1 (p < 0.01). When examining fast breathing for age, more children in ITIP3 than in ITIP1 met very fast breathing criteria. Among those less than 12 months of age, 4.7% (2/43) had very fast breathing in ITIP3’s FBP cohort compared to 2.6% (10/392) in ITIP1 (p = 0.34). Among those older than 12 months of age, 22.2% (8/36) had very fast breathing in ITIP3’s FBP cohort compared to 4.4% (32/734) in ITIP1 (p < 0.01). There was a non-significantly higher percentage of fevers in ITIP3’s FBP cohort (39.2%, 31/79) than in ITIP1 (29.8%, 335/1126; p = 0.08).

Table 2 ITIP3 fast-breathing pneumonia cohort and ITIP1 baseline characteristics
Full size table

A slightly higher percent of children in ITIP1 received age-appropriate numbers of pentavalent and pneumococcal conjugate vaccines than in ITIP3’s FBP cohort (54.4% in ITIP3 for both vaccines versus 57.6% for pentavalent vaccine (p = 0.58) and 57.5% for pneumococcal conjugate vaccine (p = 0.59) in ITIP1, Appendix S1). Overall hospitalizations were higher in ITIP3’s FBP cohort (11.4%, 9/79) compared to ITIP1 (8.7%, 98/1126), but the difference was not statistically significant (p = 0.54, Appendix S2). One child was re-hospitalized in ITIP3 and four children were re-hospitalized in ITIP1 (data not shown).

In ITIP3’s FBP cohort, TF was 4.1% (3/73) compared to 4.5% (25/552) in ITIP1’s amoxicillin treatment group (p > 0.99) (Table 3). In contrast, 7.6% (41/543) of children in ITIP1’s placebo group had TF when compared to ITIP3’s FBP cohort (p = 0.28). There were no significant differences in rates of those clinically cured by Day 14 between ITIP3’s FBP cohort (95.8%) and either ITIP1’s amoxicillin (96.7%; p = 0.73) or ITIP1’s placebo (98.1%; p = 0.19) groups (Table 4). There were no deaths in ITIP3’s FBP cohort by the Day14 assessment nor in ITIP1’s clinical trial (over the entire observation period). Post-Day 14, there were 2 deaths recorded in ITIP3’s FBP cohort, one child with danger sign pneumonia and another child with pulmonary tuberculosis and severe acute malnutrition. The child with danger sign pneumonia who died was clinically cured based on clinical data on Day 14 but died on Day 22. The child with pulmonary tuberculosis and severe acute malnutrition was clinically not cured based on clinical data on Day 14. There was no follow-up (including assessment of deaths) in ITIP1’s clinical trial post-Day 14.

Table 3 ITIP3 fast-breathing pneumonia cohort and ITIP1 treatment failure ratesa
Full size table
Table 4 ITIP3 fast-breathing pneumonia cohort and ITIP1 clinical cure ratesa
Full size table

Discussion

Many children with pneumonia have comorbidities, and these children can be at high risk for pneumonia mortality. In our ITIP3 study population, the number of children with FBP and one or more comorbidities was low (n = 79), with HIV and malaria the most common comorbidities. When comparing children in ITIP3’s FBP cohort to those in our ITIP1 FBP clinical trial, TF by Day 6 was not significantly different among children in ITIP3; and neither were clinical cure rates by Day 14. Children with FBP who were excluded from the concurrent ITIP1 clinical trial due to comorbidities did not fare worse in our prospective observational study. Because this current analysis focused only on children with FBP rather than more severe pneumonia, it is likely that we were able to identify children early in their disease course. Another possibility is that because fast breathing is a non-specific sign, some proportion of these children had a primary process other than pneumonia.

These results may be explained in part by the systematic and comprehensive risk assessment done at the first point of care for study participants. For example, while HIV is a significant risk factor for mortality, HIV screening is inconsistently done in practice [11,12,13]. However, in ITIP3, children were systematically tested for HIV such that those with HIV were identified promptly at the first point of care. While such children still received less structured care consistent with real-world clinical management, outcomes were comparable to those of lower risk children without comorbidities in ITIP1 who received more structured care and closer follow-up typical of a clinical trial. Additionally, other signs of illness such as fever and difficult breathing were identified early among these children, with many of them receiving care within 2 to 3 days of initial sign of illness. These results may suggest that if higher risk children seek care early, and risk is systematically and appropriately assessed (e.g., pulse oximetry, hemoglobin, HIV and malaria testing) at the first point of care, nearly all can achieve clinical cure by Day 14, similar to those without comorbidities [4, 14, 15].

One important caveat in this study is that most of the study population was recruited from an outpatient clinic at BDH whose catchment area is urban or peri-urban Lilongwe, and we were not able to account for children who went directly to KCH, were referred to KCH from a rural health center, or did not access care. Thus, there may have been selection bias if children with more severe cases of pneumonia bypassed BDH. Another limitation included enrollment to ITIP3’s FBP cohort being curtailed when ITIP1’s FBP trial stopped enrolling. This led to enrollment of only a small number of children with FBP, resulting in high uncertainty around estimates which was further compounded by missing outcome data. These small numbers may limit generalizability to all children with comorbidities; they also may reflect that children with pneumonia and comorbidities may progress to more severe disease faster.

We also focused on selected common, major comorbidities, and not all potential comorbidities, such as a history of being born preterm or low birthweight. Other limitations were different treatment regimens used, outcome assessments at different timepoints, and the lack of structured follow-up in ITIP3 compared to ITIP1 (study visits were more frequent for ITIP1 than ITIP3 during the first 14 days), which limited the information available regarding clinical course and made direct comparisons difficult. These structural differences were intentional because ITIP1 was a clinical trial with randomized treatments and close follow-up and ITIP3 was an observational study using standard of care treatments with some additional follow-up. We applied the relevant WHO management guidelines and timepoints for TF assessment specific to each ITIP1 and ITIP3 group. We were unable to apply the same TF criteria to both groups given the guidelines themselves are different. We chose to apply WHO guidelines for diagnosing and treating pneumonia to these groups because it was important for our results to be representative of real-world care, and these are the guidelines that healthcare providers use during routine care conditions. Finally, this is a complete case analysis and other more extensive sensitivity analyses were not performed due to the exploratory and observational nature of the study.

Conclusions

In conclusion, clinical cure rates by Day 14 were high in ITIP3’s FBP cohort. Clinical cure rates by Day 14 were similar between ITIP3 children excluded from ITIP1 and those in ITIP1 treated similarly. These children with FBP were not at high risk for mortality despite having comorbidities. If risk factors are promptly identified and relevant clinical guidelines are followed, overall outcomes among children with FBP and comorbidities can be comparable to those among low-risk children with FBP and no comorbidities. Systematic risk assessment coupled with appropriate management and referral is effective in limiting adverse outcomes in children with FBP and comorbidities in this setting.