The Epidemiology of Meconium Aspiration Syndrome: Incidence, Risk Factors, Therapies, and Outcome

METHODS

Data on infants with MAS were obtained from the database of the Australian and New Zealand Neonatal Network (ANZNN). This is a voluntary collaboration between all 28 level III NICUs that was established in 1994 to monitor the care of high-risk newborns. A minimum data set has been developed, with any infant satisfying any of the following criteria being included in the database: (1) need for assisted ventilation either through an endotracheal tube or via nasal/nasopharyngeal continuous positive airway pressure (nCPAP), (2) need for major surgery, (3) gestation <32 weeks, and (4) birth weight <1500 g. Data are supplied to ANZNN in de-identified form by the individual units and are checked for errors. For each infant, a primary respiratory diagnosis is entered.

For this study, the ANZNN database for the years 1995–2002 was searched for infants >35 weeks' gestation with a primary respiratory diagnosis of MAS. The ANZNN defines MAS in a meconium-stained infant as respiratory distress within 12 hours of age, displaying symptoms such as hypoxia, tachypnea, gasping respirations, and signs of underlying asphyxia, with a chest radiograph showing overexpansion of the lungs with widespread coarse infiltrates.²⁶ It is important to note that the diagnosis of MAS is ascribed by the clinicians responsible for supplying data to the ANZNN database. This analysis focused on cases of MAS requiring intubation and mechanical ventilation (MAS_INT). Infants with MAS treated with nCPAP only were excluded; such infants, although meconium-stained, may have other diagnoses as the predominant cause of respiratory distress, for example, transient tachypnea of the newborn.^{27, 28} All of the available antenatal and intrapartum data and information pertaining to severity of the disease and medical management, were extracted from the identified records. The presence of intrauterine growth retardation was noted, defined as birth weight less than the 10th percentile derived from Australian normative data.²⁹ For infants who died, the cause of death was classified as being respiratory or nonrespiratory (ie, directly related or not related to the aspiration of meconium) based on the International Classification of Diseases code numbers and additional diagnostic information entered in the ANZNN database.

Information regarding the total birth population in Australia and New Zealand was also obtained for the years 1995–2002. Data were gathered on the total numbers of live births and the incidence of potential risk factors for MAS that could be matched to the ANZNN MAS_INT cohort. These factors included maternal ethnicity, gender, place of birth, mode of delivery, gestational age, and condition at birth as indicated by 5-minute Apgar score. Australian data were obtained from the annual reports of the Australian Institute of Health and Welfare.³⁰ Where variables were reported in terms of the number of confinements, the same percentage was used to project the number of live births. Variables treated in this manner were indigenous births, home births, delivery by cesarean section, and instrumental deliveries. Where data were stated to be missing or incomplete from some states or territories (eg, the classification of cesarean section as “emergency” or “elective”), the national average and previous trends were used to calculate the totals.

For New Zealand, the total number of live births for each year and data on maternal ethnicity and infant gender were obtained from electronic documents made available by Statistics New Zealand.³¹ Data on type of delivery between 1999 and 2002 and numbers of instrumental vaginal deliveries between 1995 and 1998 were obtained from the Ministry of Health reports on maternity (eg, refs 32 and 33). Numbers of emergency and elective cesarean section deliveries were not separately reported in the years 1995-1998, and estimates have been made from subsequent trends. No reliable data on home births in New Zealand were available. Summary statistics on gestational age and 5-minute Apgar score were reported differently from Australian data and could not be included in the overall data set. For both countries, numbers of infants born in tertiary hospitals were obtained from the annual reports of the ANZNN.²⁶

Summary statistics were generated for all variables in the MAS data set. The duration of intubation and nCPAP were summed to give duration of respiratory support. No assumptions were made about the distribution of continuous variables. For binary and continuous variables, simple comparisons were made with χ² and Mann-Whitney tests, respectively; longitudinal trends in these variables were examined using the χ² test for trend and linear regression, respectively. For all tests, a P < .05 was considered statistically significant. Within the MAS_INT cohort, predictors of duration of respiratory support were sought using multiple linear regression analysis, with addition of all of the relevant variables to the model in a stepwise manner. Binary variables were entered as 1 or 0. Multivariate analysis could not be applied to the entire birth population because of lack of access to specific data from each individual without MAS_INT.

RESULTS

Between 1995 and 2002, 1061 infants were admitted to a level III NICU in Australia or New Zealand with MAS_INT, with the number ranging between 109 and 150 annually. An additional 107 infants with MAS treated with nCPAP only were excluded along with 2 infants with a primary diagnosis of MAS in whom no respiratory support was required. Median gestation for infants with MAS_INT was 40 weeks (interquartile range: 39–41 weeks), median birth weight was 3400 g (interquartile range: 3000–3700 g), and 52.4% were male. Overall, 42% of infants with MAS_INT were born in a tertiary hospital, with the proportion being higher in New Zealand than Australia (64% vs 31%; P < .001).

During the study period, there were 2490862 live births in the 2 countries, giving an overall rate of MAS_INT of 0.43 per 1000 live births. In Australia, there were 894 infants with MAS_INT in 2038666 live births and in New Zealand, 167 cases in 452196 live births, yielding rates of MAS_INT of 0.44 and 0.37 per 1000 live births, respectively. Combined longitudinal data from both countries revealed an overall decrease in the incidence of MAS_INT to a low of 0.35 per 1000 in 2002 (P = .0087). Analysis by country shows a steady decrease in incidence in Australia between 1995 and 2002, with considerable fluctuation but no clear trend in incidence in New Zealand.

Figure 1 shows the rate of MAS_INT per 1000 live births at each week of gestation (Australian data only). Between 36 and 40 weeks, the rate varied between 0.24 and 0.46 per 1000, with the lowest rate occurring at 38 weeks, and rose thereafter to a high of 1.42 per 1000 at 42 weeks and beyond. Thirty-four percent of the MAS_INT cohort were at >40 weeks' gestation and 6.5% at >41 weeks' compared with 16% and 2.0% of the total birth population (P < .001 in both cases). Figure 2 shows the trends in incidence of MAS_INT by gestational age grouping during the study period. A decrease in incidence was noted over time at gestations >40 weeks but not at gestational ages between 36 and 40 weeks.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Incidence of MAS_INT according to gestational age. Plot of incidence of MAS_INT at each week of gestation; Australian data only. See text for comparisons between gestational age groups.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2

Incidence of MAS_INT according to year in different gestational age groups: plot of incidence of MAS_INT per 1000 live births for gestations 36 to 40 weeks and for gestations beyond 40 weeks; Australian data only. ^a Decrease in incidence between 1995 and 2002 for infants beyond 40 weeks' gestation (P = .010), with no discernible trend for 36 to 40 weeks. □, ≥41 weeks; ▵, 36 to 40 weeks.

Antecedents of MAS_INT, with analysis by year and by country, are shown in Table 1. More than half of the infants were noted to have fetal distress prompting obstetric intervention; 70% of these went on to deliver by cesarean section or assisted vaginal delivery. Cesarean section was the mode of delivery for 42% of the MAS_INT infants overall. Fully 9.7% of the entire MAS_INT cohort were delivered by cesarean section without labor, preceded by the recognition of fetal distress in 77% of cases and of intrauterine growth restriction in an additional 6%. The proportion of assisted vaginal deliveries decreased during the study period. The number of MAS_INT infants with 5-minute Apgar score <7 and the number requiring intubation in the delivery room also decreased. Median Apgar scores at 5 minutes were lower among infants born in New Zealand (6 vs 7; P = .0012), and intubation in the delivery room for resuscitation was more commonly required (59% vs 43%; P < .001).

The impact of selected antecedents on the risk of MAS_INT, using combined data from all years, is shown in Fig 3. There was a striking association between low 5-minute Apgar score and risk of MAS_INT, with an odds ratio of 52. Among the demographic variables, infants born to Pacific Islander mothers in New Zealand and Aboriginal/Torres Strait Islander mothers in Australia were ∼3 times more likely to develop MAS_INT, whereas infants born to Maori mothers had no increased risk. Gender of the infant was not a significant factor. All forms of assisted delivery increased the risk of MAS_INT to some extent, with infants undergoing emergency cesarean section being at the highest risk, almost sixfold higher than other modes of delivery. Planned home birth was associated with a 2.7-fold increase in risk of MAS_INT.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3

Risk factors for MAS_INT within the total birth population: plot of odds ratios (⋄) and 95% confidence interval (← and →) for selected risk factors for MAS_INT. Odds ratio (95% confidence interval) also indicated as text. ^a Australian data only; ^b New Zealand data only. TSI indicates Torres Strait Islander; “All CS,” all cesarean-section deliveries.

Figure 4 shows longitudinal trends in odds ratios for selected risk factors during the study period. There was considerable fluctuation in the odds ratio for most risk factors, with no clear trend over time. The risk for infants born to Aboriginal/Torres Strait Islander mothers did, however, increase markedly from 1998, such that in 2002 the risk of MAS_INT was >7 times that of the nonindigenous population in Australia. Both emergency cesarean section and birth in a tertiary center showed a downward trend over the study period but continued to be associated with an increased risk of MAS_INT. The odds ratio for an assisted vaginal delivery decreased rapidly and was <1 after 1997.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4

Longitudinal trends in odds ratio for selected risk factors. Plot of odds ratio by year for ethnicity (A), place and mode of delivery (B), and 5-minute Apgar score (C) are shown. Odds ratios for Aboriginal/Torres Strait Islander (TSI) are specific for Australia, and those for Pacific Islander are specific for New Zealand. “Emerg CS” indicates cesarean section in labor; “Assist vag,” assisted vaginal delivery (defined in Table 1); other abbreviations as per Fig 3. ^a Increase in odds ratio for MAS_INT among Aboriginal/TSI during the study period (P = .0091). ^b Decrease over time in odds ratio for birth in a tertiary center and assisted vaginal delivery (P = .038 and 0.011, respectively). A, □, Aboriginal/TSI; ▵, Pacific Islander; ⋄, Maori. B, ▵, All CS; □, Emerg CS; ○, Assist vag; ⋄, tertiary birth. C, □, 5-min Apgar <7.

With the knowledge of the antecedents of MAS_INT in the birth population overall, we sought to identify changes in the demography of live births that might explain the decrease in incidence of MAS_INT in the latter years of the study period. Compared with 1995, in 2002, there were fewer deliveries beyond 41 weeks' gestation (1.6% vs 2.8%; P < .001) and fewer infants with a 5-minute Apgar score <7 (1.4% vs 1.7%; P < .001). These factors combined account for 62% of the reduction in MAS incidence noted in this time period.

Table 2 shows the therapies received by infants with MAS_INT. More than half of infants overall received either surfactant, HFV or iNO, with upward trends in the use of all of these therapies during the study period. Infants receiving ≥1 of surfactant, HFV, or iNO were intubated for a median of 5 days and required respiratory support for a median of 6 days, compared with 2 and 3 days, respectively, for infants who received none of these therapies. Only 1.1% of the MAS_INT cohort required treatment with extracorporeal membrane oxygenation; all infants treated with this therapy survived.

Table 3 shows the complications and outcomes for the ANZNN MAS_INT cohort. Although no clear trend was discernible for the whole study period, the incidence of air leak seemed to decrease in 2001–2002, being significantly lower for these 2 years compared with all of the other years (5.3% vs 10.8%; P = .013). Infants who developed air leak required respiratory support for a median of 5 days compared with 3.2 for those without this complication. In the MAS_INT cohort overall, the duration of intubation and of respiratory support remained relatively constant during the study period; however, the duration of oxygen therapy and the length of hospital stay increased. The overall mortality was 6.6% (70 of 1061), with the majority of these deaths being because of perinatal asphyxia (n = 33) or congenital anomalies (n = 6). Death directly attributable to aspiration of meconium occurred in 24 infants, equivalent to 2.5% of the total MAS_INT cohort and 0.96 per 100000 live births. Mortality rates seemed to remain constant during the study period. Of the 24 infants whose death was directly related to MAS, 7 (29%) were in relatively good condition at birth, not requiring immediate intubation, and with a 5-minute Apgar ≥7. Air leak occurred in 42% of MAS-related deaths compared with 13% of infants dying from other causes (P = .037) and 8.6% of survivors (P < .001). Infants dying of MAS were more likely to have received surfactant, HFV, or iNO (P = .037), but 5 infants received none of these therapies.

Forty-nine infants received home oxygen, with requirement for this therapy increasing during the study period. These infants had a longer duration of respiratory support, with a median of 7 days, compared with 3.8 days for infants who did not require home oxygen (P < .001). They were also more likely to receive surfactant, HFV, or iNO (P = .026), although 16 infants received none of these therapies.

In multiple linear regression analysis, duration of respiratory support in MAS_INT survivors was found to be predicted by the following binary variables: cesarean section in labor (coefficient: 1.33; P = .003), fetal distress (coefficient: 1.05; P = 0.012), indigenous Australian or Pacific Islander ethnicity (coefficient: 1.38; P = 0.019), and male gender (coefficient: 0.82; P = 0.040). Gestational age, intrauterine growth restriction, and 5-minute Apgar score did not independently predict duration of respiratory support when added to the model.

DISCUSSION

In this study we investigated the epidemiology of MAS in a large cohort of ventilated infants, with reference to available data for the entire birth population during the same period. We found that during the period 1995–2002, there was a reduction in the incidence of MAS_INT to <0.40 cases per 1000 live births, a strong association of MAS_INT with low 5-minute Apgar score, a clear increase in risk in Pacific Islander and indigenous Australian pregnancies, and a similar association with advanced gestation to that noted previously. A heightened risk of MAS_INT was also noted after planned home birth. Our data show a median duration of respiratory support of 4 days, with 31%, 19%, and 30% of cases receiving exogenous surfactant, HFV, and iNO, respectively. Mortality directly attributable to MAS was 2.5% in the MAS_INT cohort.

A strength of this study lies in the large numbers of infants with MAS_INT in the ANZNN cohort (1061 infants), in whom a relatively large individual data set has been assembled. In addition, ANZNN data entry allows only 1 primary respiratory diagnosis, with MAS being a clearly defined and distinct entity. This allows exclusion of meconium-stained infants with minimal evidence of parenchymal disease in whom the main diagnosis is PPHN. We have also had the advantage of obtaining data from large perinatal data registries in both countries, meaning that the findings in the ANZNN cohort can be placed in the context of the total population of 2490862 live births in which the cases of MAS_INT occurred.

The low incidence of MAS_INT and its apparent reduction during the study period confirm that modern obstetric practices can, for the most part, interrupt the chain of events that results in significant aspiration of meconium. The reduction in incidence cannot be ascribed to more zealous delivery room management, in particular, tracheal suctioning. Indeed, in the latter part of the study period, routine intubation of the trachea after MSAF was largely abandoned in Australia and New Zealand, contemporaneous with the publication of a randomized, controlled trial finding no benefit from this intervention in the vigorous infant.¹⁵ The decrease in incidence parallels the findings of Yoder et al,⁸ who, in selected birth cohorts from the same time period, found a reduction in risk of MAS, attributed to avoidance of postmaturity and more aggressive management of suspected fetal distress. The reduced incidence of MAS_INT over time in the present study would seem to relate to similar factors, with advanced gestation and the proportion of infants having a 5-minute Apgar score <7 being seen to decrease from 1995–2002.

Our data analysis allowed the computation of odds ratios for a number of potential risk factors for MAS_INT in the total birth population. In Australian live births, we noted a very strong association with a 5-minute Apgar score <7, with an overall odds ratio of 52. This is, to our knowledge, the first population-based study to quantify the risk of MAS in relation to condition of the infant at birth. The implication is that ∼1.3% of all live-born infants with a 5-minute Apgar score <7 will subsequently be ventilated with MAS compared with 0.025% of those with a 5-minute Apgar score ≥7. Other cohort studies have documented the importance of condition at birth in determining the risk of MAS developing in the context of MSAF. For this latter association, Wiswell et al,¹⁵ using stepwise linear regression, reported an odds ratio of 21; others have found the risk to be approximately fourfold.^{8, 12} There remains some dispute about how this effect is mediated; the presence of fetal distress requiring obstetric intervention in >50% of the ANZNN MAS_INT cohort suggests exaggerated fetal respiration and in utero aspiration of meconium as a possible mechanism.

The ethnology of MAS has been explored in this study, with the finding of a 3 times greater risk of MAS_INT in Pacific Islanders and indigenous Australians and, in the latter group, an apparent upsurge in risk in 2001 and 2002. These findings complement those of other investigators in identifying ethnic groups of heightened risk of MSAF and MAS. Black Americans and Africans have been found to have a risk of MSAF 1.5 to 2.6 times that of other ethnic groups,^{7, 10, 17, 19} the explanation for which may lie in earlier attainment of fetal maturity and is not related to longer gestation.¹⁹ Pacific Islanders in New Zealand have a 1.8-fold increase in risk of MSAF.¹⁸ There are no published data regarding risk of MSAF or MAS in indigenous Australians. Our population-based estimate of the risk of MAS in Pacific Islanders (odds ratio: 2.7) is similar to that noted previously within a single hospital (odds ratio: 3.6)²⁰ and cannot be entirely explained by the increase in the risk for MSAF noted in this group. At least within the MAS_INT cohort, we could not identify any other possible explanatory factors, with fetal distress, Apgar score, and gestation being equivalent among Pacific Islanders, indigenous Australians, and others (data not shown). The mechanism by which ethnicity imparts additional risk of MAS requires additional study. We found, as have others,¹⁸ no increase in the risk of MAS among the New Zealand Maori population.

The risk of MAS_INT among live-born infants was increased beyond 40 weeks' gestation, with a 3.4-fold increase in risk at >41 weeks. Only 1 other population-based study has examined the relationship between gestation and risk of MAS, reporting an odds ratio of 2.9 for the risk of MAS beyond 41 weeks.²⁴ The propensity to develop MAS at advanced gestation would seem to be related both to a higher risk of MSAF^{10, 17, 21–23} and an increased likelihood that MAS will occur in the presence of MSAF,^{8, 11, 23} independent of other risk factors, such as fetal distress and low Apgar scores.⁸

We found that place of birth was a predictor of MAS_INT, with a higher risk for infants born in tertiary centers (odds ratio: 1.65) and for planned home births (odds ratio: 2.74). This latter finding is in contrast to the reported outcome for planned home birth in Canada, where no apparent increase in the risk of either MSAF or MAS was noted.³⁴ We also found cesarean delivery to be associated with MAS_INT, as has been documented previously in a smaller birth cohort.¹⁶ The need for cesarean delivery is also known to be associated with a higher risk of subsequent MAS in the setting of MSAF.^{11, 12, 15} Fetal growth restriction has been linked previously to MAS,²⁴ and we found a doubling of risk for MAS_INT in growth-restricted infants. Most infants with MAS_INT were, however, well-grown, with average birth weight being between the 25th and 50th percentile for gender and gestation.²⁹

A weakness of our investigation is that within the total birth population it was not possible to ascertain the incidence of MSAF, a requisite intermediary in the pathogenesis of MAS. This limits, to some degree, our capacity to distinguish whether the risk factors we defined within the population exert their effect by increasing the rate of MSAF, the risk of MAS in the presence of MSAF, or both. Similarly, for this study we did not have access to data from each individual in the total birth population and, thus, can perform only univariate data analysis. In the absence of multivariate regression analysis, it is not possible to define clearly whether the heightened risks of MAS_INT related to ethnicity and gestation are subsumed in a final common pathway of fetal distress or low Apgar score. These questions clearly need additional investigation in population-based epidemiologic studies.

Despite limited data to support the use of HFV and exogenous surfactant in MAS, both of these treatments were applied with increasing frequency over the study period. Anecdotally, HFV in the form of oscillatory or jet ventilation has been found to be effective in supporting infants with MAS^{35, 36} and may be a contributing factor to the low requirement for extracorporeal membrane oxygenation in the ANZNN MAS_INT cohort. In a national survey of neonatal units in continental France, Nolent et al³⁷ found high-frequency oscillatory ventilation to be used for MAS in 27 of 29 units; a similar situation exists in Australia and New Zealand (P.A.D., unpublished data).

Of the trials of bolus surfactant therapy published during the study period,^{38, 39} only that of Findlay et al³⁸ show clear benefits beyond an improvement in oxygenation. It is, therefore, somewhat surprising that exogenous surfactant therapy has found such a prominent place in the treatment of MAS_INT. In the latter years of the study period, exogenous surfactant was administered to ∼40% of all infants with MAS_INT. No data are available in this study concerning surfactant dose and retreatment.

Among the ANZNN MAS_INT cohort, 51% of infants received ≥1 HFV, iNO, and exogenous surfactant, with a documented increase in this proportion from 1996 to 2002. The uptake of these newer therapies over the study period seems to have had no effect on the duration of intubation, and the duration of oxygen therapy and length of hospital stay have increased in the MAS_INT cohort overall. These findings emphasize the supportive rather than curative nature of these therapies in infants with MAS_INT.

The mortality risk for ventilated infants with MAS in Australia and New Zealand remains significant. Overall mortality was 6.6%, with deaths directly attributable to the respiratory illness occurring in 2.5% of the entire MAS_INT cohort and 0.96 per 100000 live births. This latter figure compares favorably with other population-based estimates of MAS-related mortality. Sriram et al⁷ found a mortality of 2.0 per 100000 live-born infants in the United States in 1998–2000; Nolent et al³⁷ report 1.03 deaths per 100000 births in France in 2000–2001, but incomplete ascertainment of MAS_INT cases in this study may limit the reliability of this estimate. These population-based estimates of mortality from MAS are generally markedly lower than those derived from hospital birth cohorts. The largest of these, studying cases of MAS in 7 hospitals in the United States during the years 1973–1987 (total birth population: 176790), found a mortality rate of 28 per 100000 live births.⁴ Another study found a mortality rate of 22 per 100000 among 85562 live births in 1 hospital in Saudi Arabia during the period 1989–1994.⁴⁰ The apparent high mortality in hospital-based studies is a reflection of the high-risk birth populations from which the estimates are made and can be presumed to also be related to the time period in which the studies were done, before the advent of a number of newer therapies that seem to have improved outcome in MAS. Other more recent studies of smaller hospital-based cohorts have shown lower mortality, including, in several instances, no MAS-related deaths.^{8, 41}

Pneumothorax remains an important complication of MAS, occurring in 9.6% of the MAS_INT cohort overall, with an apparent reduction to ∼5% in 2001–2002. Other recent studies have reported an incidence of pneumothorax in intubated infants of 8%,³⁸ 14%,⁴² and 20%,⁴³ although this last study selected infants with more severe disease. Despite the advances in respiratory support for MAS, pneumothorax remains an important indicator of disease severity, being present in 42% of infants who ultimately died of MAS in the ANZNN cohort.

Durations of intubation and respiratory support did not alter throughout the study period, although more infants received nCPAP either before or after mechanical ventilation. Disconcertingly, the duration of oxygen therapy and length of hospital stay increased during the study period, reaching a median of 8 and 17 days, respectively, in 2002. These figures emphasize the considerable burden that MAS continues to place on afflicted infants, their families, and the health care system.

CONCLUSIONS

We found in this study that the incidence of MAS requiring intubation in 2 countries in the developed world is low and seems to be decreasing. The risk of MAS_INT is significantly greater in the presence of fetal distress and low Apgar score, as well as Pacific Islander and indigenous Australian ethnicity. MAS is now frequently treated with newer therapeutic modalities, such as HFV, iNO, and exogenous surfactant, but the duration of ventilation and oxygen therapy have not been improved as a result. Mortality for MAS_INT is low, but there remains a significant risk of pneumothorax.