Prepared select statements don't reflect table alterations without Cluster re-instantiation in Cassandra

CassandraPrepareStatement.java

@Test
public void reprepareStatement() {
    Cluster cluster = Cluster.builder().addContactPoints("localhost").build();
    Session session = cluster.connect("test");

    Statement create = createTable("test")
                .addPartitionKey("id", DataType.text())
                .addColumn("initial_column", DataType.text())
                .ifNotExists();

    session.execute(create);

    PreparedStatement prepared = session.prepare("select * from test");
    ResultSet result = session.execute(prepared.bind());
    logger.info("Columns before alter table: {}", getColumnDefinitions(result));
    // Columns before alter table: [id, initial_column]

    session.execute(alterTable("test").addColumn("new_column").type(DataType.text()));

    result = session.execute(prepared.bind());
    logger.info("Columns after alter table: {}", getColumnDefinitions(result));
    // Columns after alter table: [id, initial_column]

    result = session.execute("select * from test");
    logger.info("Columns after unprepared execute: {}", getColumnDefinitions(result));
    // Columns after unprepared execute: [id, initial_column, new_column]

    prepared = session.prepare("select * from test");
    result = session.execute(prepared.bind()); // Cassandra Warning: "Re-preparing already prepared query select * from test. Please note that preparing the same query more than once is generally an anti-pattern and will likely affect performance. Consider preparing the statement only once."
    logger.info("Columns after re-prepare: {}", getColumnDefinitions(result));
    // Columns after re-prepare: [id, initial_column]

    session.close();
    session = cluster.connect("test");
    prepared = session.prepare("select * from test"); // Cassandra Warning: "Re-preparing already prepared query select * from test. Please note that preparing the same query more than once is generally an anti-pattern and will likely affect performance. Consider preparing the statement only once."
    result = session.execute(prepared.bind());
    logger.info("Columns recreating session via re-connect on cluster: {}", getColumnDefinitions(result));
    // Columns recreating session via re-connect on cluster: [id, initial_column]


    session.close();
    cluster.close();
    session = Cluster.builder().addContactPoints("localhost").build().connect("test");
    prepared = session.prepare("select * from test");
    result = session.execute(prepared.bind());
    logger.info("Columns after recreating session via re-instantiating cluster: {}", getColumnDefinitions(result));
    // Columns after recreating session via re-instantiating cluster: [id, initial_column, new_column]

    session.execute(dropTable("test"));
}

private List<String> getColumnDefinitions(ResultSet result) {
    return result.getColumnDefinitions()
            .asList()
            .stream()
            .map(definition -> definition.getName())
            .collect(Collectors.toList());
}