Excerpts from the Literature
Articles of interest compiled by the editorial board.
Use of Inhaled Nitric Oxide in Preterm Infants Summarized
Praveen Kumar and COMMITTEE ON FETUS AND NEWBORN. Pediatrics 2014;133;164 (UG)
Due to its vasodilating effects, nitric oxide (NO) is used in the treatment of full-term and late-preterm infants with persistent pulmonary hypertension of the newborn and hypoxemic respiratory failure. Several randomized controlled trials have evaluated the role NO in the management of preterm infants with varying results. In this paper the authors summarize the existing evidence for the use of inhaled nitric oxide in preterm infants and provide guidance regarding its use in this population. The authors reviewed the literature on the use of NO in preterm infants with respiratory failure and in preterm infants to improve the rate of survival without bronchopulmonary dysplasia (BPD). Among other findings, their meta-analysis indicated that neither rescue nor routine use of iNO improves survival in preterm infants with respiratory failure (Evidence quality, A; Grade of recommendation, strong); evidence does not support treating preterm infants who have respiratory failure with iNO for the purpose of preventing/ ameliorating BPD, severe intraventricular hemorrhage, or other neonatal morbidities (Evidence quality, A; Grade of recommendation, strong); the incidence of cerebral palsy, neurodevelopmental impairment, or cognitive impairment in preterm infants treated with iNO is similar to that of control infants (Evidence quality, A). They also concluded that there are limited data and inconsistent results regarding the effects of iNO treatment on pulmonary outcomes of preterm infants in early childhood.
Validity of establishing pediatric reference intervals based on hospital patient data: A comparison of the modified Hoffmann approach to CALIPER reference intervals obtained in healthy children. (JS)
Shaw JLV, Cohen A, Konforte D, Binesh-Marvasti T, Colantonio DA, Adeli K. Clin Biochem (2013) (2014) 47(3):166-72
Establishing reference intervals in a healthy population of individuals requires an exorbitant amount of work. Two considerable challenges include the need to carefully define and verify “healthy” (often on an analyte-by-analyte basis) and the need for a minimum of 120 individual specimens per partition (e.g., age, gender, menopausal status, smoking status, etc.). These specific challenges are exponentially more difficult in a pediatric population. Biological samples are often not obtained from children unless pathology is suspected and total blood collection amounts are limited by patient size.
Methods to establish reference intervals based on data from hospitalized patients have been suggested previously, most notably the “Hoffmann approach” first published in 1963 (JAMA (1963)185:150-9, Clin Chim Acta (2009)405:43-48). The benefit of these approaches is that numerous samples are readily available, thereby circumventing the types of challenges outlined above. The Hoffmann approach assumes two important factors. First, it assumes data from hospitalized patients forms a Gaussian distribution. Second, hospitalized patients are assumed to represent normal individuals. In 2008, Soldin et al. recommended that reliable intervals may be generated using this type of statistical approach if at least 50% of individuals in the reference population are healthy (Clin Biochem (2008)41:937-42).
The goal of the Shaw et al. publication was to determine the validity of reference intervals determined using inpatients (a modified “Hoffmann approach”) as compared to those determined from the healthy CALIPER (Canadian Laboratory Initiative in Pediatric Reference Intervals) repository. The CALIPER initiative is a collaborative project that aimed to recruit over 2000 healthy children from 0 to 18 years of age. They included 12 analytes measured using the Vitros 5600 in their analysis (albumin, alkaline phosphatase, ALT, AST, calcium, cholesterol, creatinine, HDL-cholesterol, iron, phosphate, triglycerides and magnesium). The number of patients used to calculate intervals varied by age, gender and analyte (range 10 to 231 after outlier removal). Hospital-based data were age- and gender-partitioned similar to previously published data using the CALIPER population. Ninety percent confidence intervals and reference change values (RCV) were calculated to compare the hospitalized and healthy intervals. Reference samples were also measured to verify the intervals determined using the Hoffmann approach (<10% outside the proposed interval, per CLSI guidelines).
The authors’ predominant finding was that intervals based on hospitalized patients were much wider than those determined using healthy individuals. No reference intervals calculated using the Hoffmann approach fell within the 90% confidence intervals as calculated using the CALIPER data. One partition for creatinine and most phosphate partitions had RCV within acceptable limits. The authors surmise that the inclusion of biological variation, which can be significant for some analytes, in the RCV calculation explains why comparisons between Hoffmann and CALIPER intervals based on RCV were more favorable than those based on confidence intervals. The reference samples overall verified the Hoffmann intervals, presumably due to the wide acceptance criteria (<10%) outlined by the CLSI guidelines.
Overall, the authors concluded that the Hoffmann statistical approach is limited in pediatrics, particularly when data originates from a tertiary care center. In general, it may be surmised that hospitalized children comprise a “sicker” population than an adult inpatient population, thus contributing to the wider intervals determined using this patient group. Due to this, use of the Hoffmann approach in pediatric populations was not endorsed by the authors. The authors suggested that outpatient clinics or community hospitals may be a better source of inpatient data if the Hoffmann approach must be used in pediatric cases. They also concluded that the CLSI guideline for verification of reference intervals may be too lenient in cases such as these.
This conclusion is similar to the findings published by Roberts et al. in a 2010 Letter to the Editor (Clin Biochem (2010)43:933-4). The comparison of intervals determined using the Hoffmann approach (Clin Biochem (2009)42:823-7) to those derived from healthy children (CHILDx sample repository) suggested limitations in using pediatric samples from inpatient or clinic settings. Comparisons of testosterone and 17-hydroxyprogesterone specifically highlighted these issues, as it is presumed that differences in intervals may be attributed to the clinical presentation of patients used for the Hoffmann approach. This is another example suggesting that use of hospitalized or clinic patients may not be appropriate to establish pediatric reference intervals and caution should be used when considering applying this statistical approach.
Trends in the Prevalence of Ketoacidosis at Diabetes Diagnosis: The SEARCH for Diabetes in Youth Study (VLP)
Dabelea D, Rewers A, Stafford JM, Standiford DA, Lawrence JM, Saydah S, Imperatore G, D'Agostino RB Jr, Mayer-Davis EJ, Pihoker C. Pediatrics. 2014 Apr;133(4):e938-45.
In this latest study by the SEARCH for Diabetes in Youth Study Group the goal was to report the prevalence of diabetic ketoacidosis (DKA) at diagnosis of type 1 and 2 diabetes. The study consisted of 7040 subjects aged 0 to 19 years, of which 5615 were diagnosed with type 1 and 1425 with type 2. The data collected spanned an 8 year period from 2002 to 2010. The criteria for reporting DKA were: if bicarbonate was <15 mmol/L, and/or a pH<7.25 (venous) or <7.30 (arterial/capillary), and/or DKA diagnosis was on the medical record.
After analyzing the data the authors found that the prevalence of DKA was higher in type 1 versus type 2 diagnosis. In type 1 patients, DKA had a prevalence of approximately 30%, and was stable throughout the time period analyzed. In type 2 patients, the prevalence was approximately 12% at the start of the time period and it progressively decreased by about 10% per year until the end of the period to 5.7%. In both types, younger patients had a higher prevalence of DKA when compared to older subjects. The data in type 1 patients was consistent with other similar studies, while the type 2 data was novel, according to the authors.
The authors also looked at some of the factors that were associated with DKA presentation. These data was obtained by questionnaire forms. The results showed that DKA was more prevalent in minority populations, lower family income and lack of private insurance. Limitations to the study included the relative short time period that could have prevented detection of small changes in DKA prevalence. Also, minority patient’s records were more likely to have missing medical information, therefore their data could have been underrepresented. It would be interesting to see in the future the inclusion of more laboratory data in the evaluation and follow-up of patients with DKA, including glucose, HgbA1C % and ketone values.