The objectives of this study were to assess the quality of 13 essential drugs commonly used in Zimbabwe; to compare the quality of drugs at selected health centres and district hospitals with the initial quality measured in the central medical stores; and to determine whether substandard drug quality, where found, is associated with specific factors: age, duration of storage and conditions at facility level, transport method or manufacturer of the drug.
This was a longitudinal study of 18 months duration. Drug batches were sampled from central medical stores. At hospitals and health centres, two samples were taken of each drug: (1) oldest batch present; and (2) same batch as taken at central medical stores.
The setting of this study was two purchasing medical stores in Zimbabwe (Harare and Bulaweyo). Health centres and hospitals from five districts represented the worst case: the farthest points of the distribution chain in low lying regions with a hot climate.
Study drugs and sample numbers
Selected drugs satisfied two criteria: (1) known or suspected quality or stability problems under adverse climatic conditions; (2) relevance for public health care (drugs on the national essential drugs list with high turnover in the public sector drug distribution system). 13 drugs were selected for study: amoxicillin capsules, ampicillin capsules, acetylsalicylic acid tablets*, doxycycline capsules, ferrous sulfate tablets*, phenoxymethylpenicillin tablets*, tetracycline capsules*, retinol tablets*, epinephrine injection*, ampicillin injection, ergometrine injection*, benzylpenicillin injection* and procaine benzylpenicillin injection*. All 13 were used at hospitals of which nine* were also used at health centres. All of the above drugs were manufactured in Zimbabwe; in addition, two drugs were procured from foreign manufacturers. 881 samples were collected. 41 samples were expired; these were excluded and considered separately. 192 samples were taken from government medical stores' stocks (CMS samples). 648 samples were taken from hospitals and health centres (facility samples). Within the facility sample group, 264 samples could be paired with CMS samples of the same batch (longitudinal series).
Main measure of drug quality
An assay value for the active ingredient of a drug was expressed as percentage of the stated content. This complied with the relevant pharmacopoeal specifications (BP, USP).
At central medical stores, 10 drugs consistently conformed to specifications. Three drugs frequently failed to comply with specifications and showed low potency: retinol tablets - 100% failures (mean assay 73.0%); ergometrine injection 65% failures (mean assay 82.4%); and ampicillin injection 20% failures (mean assay 96.7%, minimum value 86.8). Facility samples showed similar substandard quality in two of these drugs: retinol tablets with 90% failures (mean assay 74.8%) and ampicillin injection with 21% failures (mean assay 96.2%, minimum value 82.7. Ergometrine injection worsened to 72% failures (mean assay 73.5%); failure rates were significantly different between manufacturers. Procaine benzylpenicillin injection showed 3% failures at facility level (mean assay 98.3%, minimum value 72.2%) although all samples passed at central medical stores. The longitudinal series indicated a significant loss of potency in two drugs: ergometrine injection (mean loss 17.1% after 4.8 months) and procaine benzylpenicillin injection (3.6% loss after 4.3 months).
Nine of the 13 drugs maintained satisfactory quality up to endpoint in the public sector health care system in Zimbabwe. Initial quality was consistently poor in retinol tablets and occasionally there was a problem in ampicillin injection, Ergometrine injection showed poor initial quality as well as serious instability, with different outcomes according to the manufacturer. Procaine benzylpenicillin injection was moderately unstable, showing potential for unacceptable loss of active ingredient within the term of shelf-life.
The consistent low potency found at facility level had clinical consequences in the case of ergometrine injection and retinol tablets (25-26% below stated content). Some of the samples of ampicillin injection and procaine benzylpenicillin showed individual low values which may compromise the efficacy of therapy.