Typhoid fever in children: a fourteen-year experience.

From 1982 to 1995, 71 children admitted in our medical center were diagnosed to have typhoid fever by culture or serology. Of the 71 children, most (83%) were aged 5-15 years. These children usually presented with fever and gastrointestinal symptoms, including abdominal pain, diarrhea, nausea or vomiting, and constipation. Hepatosplenomegaly was the most common physical sign observed and abdominal tenderness ranked the second. Thrombocytopenia occurring in 9 patients (13%) was the most common mode of complication. Other complications included intestinal perforation (3%), rectal bleeding (3%), ascites or pleural effusion (4%), and meningitis (1%). The incidence of complications tended to be higher among children 5 years of age or older (p = 0.31). Most patients responded well to appropriate antimicrobial therapies. There was no mortality. Relapse was observed in two children, although both had received 10 days of chloramphenicol therapy. The clinical isolates of Salmonella typhi were susceptible in vitro to all the antibiotics tested, including chloramphenicol, which, however, showed a higher MIC90 level than other drugs tested. In conclusion, there were age-specific differences of typhoid fever in children in terms of the incidence and morbidity and antibiotic resistance of S. typhi has not been a problem in this area at least up to 1995.

Robust radial basis function neural networks.

Function approximation has been found in many applications. The radial basis function (RBF) network is one approach which has shown a great promise in this sort of problems because of its faster learning capacity. A traditional RBF network takes Gaussian functions as its basis functions and adopts the least-squares criterion as the objective function, However, it still suffers from two major problems. First, it is difficult to use Gaussian functions to approximate constant values. If a function has nearly constant values in some intervals, the RBF network will be found inefficient in approximating these values. Second, when the training patterns incur a large error, the network will interpolate these training patterns incorrectly. In order to cope with these problems, an RBF network is proposed in this paper which is based on sequences of sigmoidal functions and a robust objective function. The former replaces the Gaussian functions as the basis function of the network so that constant-valued functions can be approximated accurately by an RBF network, while the latter is used to restrain the influence of large errors. Compared with traditional RBF networks, the proposed network demonstrates the following advantages: (1) better capability of approximation to underlying functions; (2) faster learning speed; (3) better size of network; (4) high robustness to outliers.

Rickets of premature infants induced by calcium deficiency. A case report.

Rickets of prematurity is not uncommon in neonatal intensive care units. Nutritional rickets in childhood is usually caused by vitamin D deficiency, but the rickets of prematurity is mainly attributable to calcium and phosphorus deficiencies. We present a premature infant with sequelae of necrotizing enterocolitis who needed prolonged administration of total parenteral nutrition (TPN), and who sustained ricketic fracture. After high calcium-fortified TPN supplementation the fracture healed well, and serum alkaline phosphatase dropped. This finding shows (1) serum calcium and phosphorus levels are of predictive value regarding rickets, (2) regular follow-ups of alkaline phosphatase levels combined with radiography in high-risk groups of premature infants are good tools for monitoring rickets, and (3) prolonged TPN administration needs to contain higher calcium and phosphorus concentrations in prematurity than in childhood.