There is only a Portuguese version of the Mozambique 2011 DHS. If (like me) you cannot read Portuguese, you can probably still understand the tables by comparing with another country's report that is in English. The tables for each topic are pretty much the same in all surveys--except for the numbers, of course!
I would advise you not to use ANY of the mm variables for an individual-level analysis. The only appropriate use of these variables is to estimate adult and maternal mortality for the entire country.
The sisters of the respondent may be quite different from the respondent. They may live (or have lived) in a different region of the country, with a different type of place of residence (urban/rural), different education, wealth quintile, etc. Even more relevant, you certainly cannot assume that the sisters had the same access to health services or use of health services as the respondent.
Remember also that the MMR in DHS reports refers to a long interval of time, typically seven years, so the median date of sisters' deaths will precede the median date of the respondent's most recent birth, often by several years.
Many people would like to use DHS data to like link maternal mortality to maternal health services. It is possible to do this with countries as units, but not, unfortunately, with individuals or even sub-national areas as units. Even with countries as units, it is difficult to establish a relationship.
I suggest that you look at DHS Analytical Study #46, available on the website. In that report we looked at the impact of maternal health care and circumstances of delivery on neonatal mortality. This relationship is much easier to analyze and might be an alternative topic for your research.