how shall i transpose rows into new columns only if duplicates appears in rows?

If you need simple, fixed column solution, you can use:

SELECT * INTO #temp FROM
(VALUES 
    (1, 1, 'EA'),
    (1, 10, 'BX'),
    (1, 100, 'CA'),
    (2, 1, 'EA'),
    (2, 10, 'RL')) T(ItemId, Pack, UOM)

SELECT ItemId, MAX([1]) Pack1, MAX([U1]) UOM1, MAX([2]) Pack1, MAX([U2]) UOM1, MAX([3]) Pack1, MAX([U3]) UOM1 FROM
(
    SELECT *,
        ROW_NUMBER() OVER (PARTITION BY ItemId ORDER BY Pack) PackNum,
        'U'+CONVERT(varchar(10), ROW_NUMBER() OVER (PARTITION BY ItemId ORDER BY Pack)) UOMNum
    FROM #temp
) T
PIVOT
(
    MAX(Pack) FOR PackNum IN ([1], [2], [3])
) PPack
PIVOT
(
    MAX(UOM) FOR UOMNum IN ([U1],[U2],[U3])
) PUOM
GROUP BY ItemId

If number of columns is not fixed, you must do it dynamically (see @gofr1) or:

DECLARE @count TABLE(N varchar(10))
    INSERT @count
    SELECT CONVERT(varchar(10), ROW_NUMBER() OVER (ORDER BY (SELECT 1)))
    FROM #temp WHERE ItemId =
    (
        SELECT TOP 1 ItemId
        FROM #temp
        GROUP BY ItemId
        ORDER BY COUNT(*) DESC
    )

DECLARE @select varchar(MAX) = STUFF((SELECT ', MAX(['+N+']) Pack'+N+', MAX([U'+N+']) UOM'+N FROM @count FOR XML PATH('')), 1, 2, '')
DECLARE @pivot1 varchar(MAX) = STUFF((SELECT ', ['+N+']' FROM @count FOR XML PATH('')), 1, 2, '')
DECLARE @pivot2 varchar(MAX) = STUFF((SELECT ', [U'+N+']' FROM @count FOR XML PATH('')), 1, 2, '')

DECLARE @sql varchar(MAX) = '
SELECT ItemId, '+@select+' FROM
(
    SELECT *,
        ROW_NUMBER() OVER (PARTITION BY ItemId ORDER BY Pack) PackNum,
        ''U''+CONVERT(varchar(10), ROW_NUMBER() OVER (PARTITION BY ItemId ORDER BY Pack)) UOMNum
    FROM #temp
) T
PIVOT
(
    MAX(Pack) FOR PackNum IN ('+@pivot1+')
) PPack
PIVOT
(
    MAX(UOM) FOR UOMNum IN ('+@pivot2+')
) PUOM
GROUP BY ItemId'

EXEC(@sql)

Dynamic SQL and some pivoting:

DECLARE @sql nvarchar(max),
        @columns nvarchar(max)

SELECT @columns = (
    SELECT DISTINCT  ','+QUOTENAME('pack' + cast(ROW_NUMBER() OVER (PARTITION BY itemid ORDER BY itemid) as nvarchar(max)))+','+
            QUOTENAME('UOM'+ cast(ROW_NUMBER() OVER (PARTITION BY itemid ORDER BY itemid) as nvarchar(max)))
    FROM #T1
    FOR XML PATH('')
)

SELECT @sql = N'
SELECT *
FROM (
    SELECT  itemid,
            Items+rn as Items,
            [Values]
    FROM (
        SELECT  itemid,
                CAST(UOM as nvarchar(max)) as UOM,
                CAST(pack as nvarchar(max)) as pack,
                CAST(ROW_NUMBER() OVER (PARTITION BY itemid ORDER BY (SELECT NULL)) as nvarchar(max)) as rn
        FROM #T1
        ) as t
    UNPIVOT (
        [Values] FOR Items in ([UOM],[pack])
    ) as up
    ) t1
PIVOT (
    MAX([Values]) FOR Items IN ('+STUFF(@columns,1,1,'')+')
) as pvt'

EXEC sp_executesql @sql

Output:

itemid  pack1   UOM1    pack2   UOM2    pack3   UOM3
1       1       EA      10      BX      100     CA
2       1       EA      10      RL      NULL    NULL

EDIT#1

There is one more way:

DECLARE @sql nvarchar(max),
        @columns nvarchar(max),
        @colsrn nvarchar(max)

;WITH cte AS (
SELECT DISTINCT cast(ROW_NUMBER() OVER (PARTITION BY itemid ORDER BY itemid) as nvarchar(max)) as rn
FROM #T1
)

SELECT @columns = (
    SELECT ',MAX('+QUOTENAME('pack' + rn) +') as '+ QUOTENAME('pack' + rn) +',MAX('+
            QUOTENAME('UOM'+ rn) +') as ' + QUOTENAME('UOM'+ rn)
    FROM cte
    FOR XML PATH('')
),
        @colsrn = (
    SELECT DISTINCT  ',CASE WHEN rn = ' + rn +' THEN '+QUOTENAME('pack')+' ELSE NULL END as [pack'+rn+'],' 
                    +'CASE WHEN rn = ' + rn +' THEN '+QUOTENAME('UOM')+' ELSE NULL END as [UOM'+rn+']'  
    FROM cte
    FOR XML PATH('')
)

SELECT @sql = N'
SELECT  itemid'+@columns+'
FROM (
    SELECT  itemid'+@colsrn+'
    FROM (
            SELECT  *,
                    ROW_NUMBER() OVER (PARTITION BY itemid ORDER BY (SELECT NULL)) as rn
            FROM #T1
            ) as t
    ) s
GROUP BY s.itemid'

EXEC sp_executesql @sql

Same output.